RomanNum3ral
/
Cybersecurity-Projects
mirror of https://github.com/CarterPerez-dev/Cybersecurity-Projects.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

checkpoint - phase 13 complete

This commit is contained in:
CarterPerez-dev 2026-02-14 02:24:37 -05:00
parent ad5a7e5f7a
commit 3fd9f77335
55 changed files with 7858 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
PROJECTS/advanced/ai-threat-detection/.env.example Normal file
View File

@ -0,0 +1,40 @@
# ©AngelaMos | 2026
# .env.example
# Application
ENV=development
DEBUG=true
HOST=0.0.0.0
PORT=8000
LOG_LEVEL=INFO
API_KEY=changeme-generate-a-real-key
# Docker Host Ports (change these if you have conflicts)
POSTGRES_HOST_PORT=16969
REDIS_HOST_PORT=26969
BACKEND_HOST_PORT=36969
# PostgreSQL
POSTGRES_DB=angelusvigil
POSTGRES_USER=vigil
POSTGRES_PASSWORD=change-this-password
DATABASE_URL=postgresql+asyncpg://vigil:change-this-password@postgres:5432/angelusvigil
# Redis
REDIS_URL=redis://redis:6379
# GeoIP (MaxMind - register at https://www.maxmind.com/en/geolite2/signup)
GEOIP_ACCOUNT_ID=
GEOIP_LICENSE_KEY=
GEOIP_DB_PATH=/usr/share/GeoIP/GeoLite2-City.mmdb
# Nginx Log Path (inside container)
NGINX_LOG_PATH=/var/log/nginx/access.log
# Pipeline Tuning
RAW_QUEUE_SIZE=1000
PARSED_QUEUE_SIZE=500
FEATURE_QUEUE_SIZE=200
ALERT_QUEUE_SIZE=100
BATCH_SIZE=32
BATCH_TIMEOUT_MS=50

72
PROJECTS/advanced/ai-threat-detection/.gitignore vendored Normal file
View File

@ -0,0 +1,72 @@
# ©AngelaMos | 2026
# .gitignore
# Planning docs
.angelusvigil/
# Environment
.env
.env.local
# Python
__pycache__/
*.py[cod]
*$py.class
*.so
.Python
build/
develop-eggs/
dist/
downloads/
eggs/
.eggs/
lib/
lib64/
parts/
sdist/
var/
wheels/
*.egg-info/
.installed.cfg
*.egg
venv/
.venv/
ENV/
env/
# Testing
.pytest_cache/
.coverage
htmlcov/
.mypy_cache/
.ruff_cache/
# IDE
.vscode/
.idea/
*.swp
*.swo
*~
# Logs
*.log
# Docker
docker-compose.override.yml
# Database
*.db
*.sqlite
*.sqlite3
# macOS
.DS_Store
# Runtime data
data/models/
data/geoip/
data/sample-logs/*.log
# MLflow
mlruns/
mlartifacts/

661
PROJECTS/advanced/ai-threat-detection/LICENSE Normal file
View File

@ -0,0 +1,661 @@
GNU AFFERO GENERAL PUBLIC LICENSE
Version 3, 19 November 2007
Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
Preamble
The GNU Affero General Public License is a free, copyleft license for
software and other kinds of works, specifically designed to ensure
cooperation with the community in the case of network server software.
The licenses for most software and other practical works are designed
to take away your freedom to share and change the works. By contrast,
our General Public Licenses are intended to guarantee your freedom to
share and change all versions of a program--to make sure it remains free
software for all its users.
When we speak of free software, we are referring to freedom, not
price. Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
them if you wish), that you receive source code or can get it if you
want it, that you can change the software or use pieces of it in new
free programs, and that you know you can do these things.
Developers that use our General Public Licenses protect your rights
with two steps: (1) assert copyright on the software, and (2) offer
you this License which gives you legal permission to copy, distribute
and/or modify the software.
A secondary benefit of defending all users' freedom is that
improvements made in alternate versions of the program, if they
receive widespread use, become available for other developers to
incorporate. Many developers of free software are heartened and
encouraged by the resulting cooperation. However, in the case of
software used on network servers, this result may fail to come about.
The GNU General Public License permits making a modified version and
letting the public access it on a server without ever releasing its
source code to the public.
The GNU Affero General Public License is designed specifically to
ensure that, in such cases, the modified source code becomes available
to the community. It requires the operator of a network server to
provide the source code of the modified version running there to the
users of that server. Therefore, public use of a modified version, on
a publicly accessible server, gives the public access to the source
code of the modified version.
An older license, called the Affero General Public License and
published by Affero, was designed to accomplish similar goals. This is
a different license, not a version of the Affero GPL, but Affero has
released a new version of the Affero GPL which permits relicensing under
this license.
The precise terms and conditions for copying, distribution and
modification follow.
TERMS AND CONDITIONS
0. Definitions.
"This License" refers to version 3 of the GNU Affero General Public License.
"Copyright" also means copyright-like laws that apply to other kinds of
works, such as semiconductor masks.
"The Program" refers to any copyrightable work licensed under this
License. Each licensee is addressed as "you". "Licensees" and
"recipients" may be individuals or organizations.
To "modify" a work means to copy from or adapt all or part of the work
in a fashion requiring copyright permission, other than the making of an
exact copy. The resulting work is called a "modified version" of the
earlier work or a work "based on" the earlier work.
A "covered work" means either the unmodified Program or a work based
on the Program.
To "propagate" a work means to do anything with it that, without
permission, would make you directly or secondarily liable for
infringement under applicable copyright law, except executing it on a
computer or modifying a private copy. Propagation includes copying,
distribution (with or without modification), making available to the
public, and in some countries other activities as well.
To "convey" a work means any kind of propagation that enables other
parties to make or receive copies. Mere interaction with a user through
a computer network, with no transfer of a copy, is not conveying.
An interactive user interface displays "Appropriate Legal Notices"
to the extent that it includes a convenient and prominently visible
feature that (1) displays an appropriate copyright notice, and (2)
tells the user that there is no warranty for the work (except to the
extent that warranties are provided), that licensees may convey the
work under this License, and how to view a copy of this License. If
the interface presents a list of user commands or options, such as a
menu, a prominent item in the list meets this criterion.
1. Source Code.
The "source code" for a work means the preferred form of the work
for making modifications to it. "Object code" means any non-source
form of a work.
A "Standard Interface" means an interface that either is an official
standard defined by a recognized standards body, or, in the case of
interfaces specified for a particular programming language, one that
is widely used among developers working in that language.
The "System Libraries" of an executable work include anything, other
than the work as a whole, that (a) is included in the normal form of
packaging a Major Component, but which is not part of that Major
Component, and (b) serves only to enable use of the work with that
Major Component, or to implement a Standard Interface for which an
implementation is available to the public in source code form. A
"Major Component", in this context, means a major essential component
(kernel, window system, and so on) of the specific operating system
(if any) on which the executable work runs, or a compiler used to
produce the work, or an object code interpreter used to run it.
The "Corresponding Source" for a work in object code form means all
the source code needed to generate, install, and (for an executable
work) run the object code and to modify the work, including scripts to
control those activities. However, it does not include the work's
System Libraries, or general-purpose tools or generally available free
programs which are used unmodified in performing those activities but
which are not part of the work. For example, Corresponding Source
includes interface definition files associated with source files for
the work, and the source code for shared libraries and dynamically
linked subprograms that the work is specifically designed to require,
such as by intimate data communication or control flow between those
subprograms and other parts of the work.
The Corresponding Source need not include anything that users
can regenerate automatically from other parts of the Corresponding
Source.
The Corresponding Source for a work in source code form is that
same work.
2. Basic Permissions.
All rights granted under this License are granted for the term of
copyright on the Program, and are irrevocable provided the stated
conditions are met. This License explicitly affirms your unlimited
permission to run the unmodified Program. The output from running a
covered work is covered by this License only if the output, given its
content, constitutes a covered work. This License acknowledges your
rights of fair use or other equivalent, as provided by copyright law.
You may make, run and propagate covered works that you do not
convey, without conditions so long as your license otherwise remains
in force. You may convey covered works to others for the sole purpose
of having them make modifications exclusively for you, or provide you
with facilities for running those works, provided that you comply with
the terms of this License in conveying all material for which you do
not control copyright. Those thus making or running the covered works
for you must do so exclusively on your behalf, under your direction
and control, on terms that prohibit them from making any copies of
your copyrighted material outside their relationship with you.
Conveying under any other circumstances is permitted solely under
the conditions stated below. Sublicensing is not allowed; section 10
makes it unnecessary.
3. Protecting Users' Legal Rights From Anti-Circumvention Law.
No covered work shall be deemed part of an effective technological
measure under any applicable law fulfilling obligations under article
11 of the WIPO copyright treaty adopted on 20 December 1996, or
similar laws prohibiting or restricting circumvention of such
measures.
When you convey a covered work, you waive any legal power to forbid
circumvention of technological measures to the extent such circumvention
is effected by exercising rights under this License with respect to
the covered work, and you disclaim any intention to limit operation or
modification of the work as a means of enforcing, against the work's
users, your or third parties' legal rights to forbid circumvention of
technological measures.
4. Conveying Verbatim Copies.
You may convey verbatim copies of the Program's source code as you
receive it, in any medium, provided that you conspicuously and
appropriately publish on each copy an appropriate copyright notice;
keep intact all notices stating that this License and any
non-permissive terms added in accord with section 7 apply to the code;
keep intact all notices of the absence of any warranty; and give all
recipients a copy of this License along with the Program.
You may charge any price or no price for each copy that you convey,
and you may offer support or warranty protection for a fee.
5. Conveying Modified Source Versions.
You may convey a work based on the Program, or the modifications to
produce it from the Program, in the form of source code under the
terms of section 4, provided that you also meet all of these conditions:
a) The work must carry prominent notices stating that you modified
it, and giving a relevant date.
b) The work must carry prominent notices stating that it is
released under this License and any conditions added under section
7. This requirement modifies the requirement in section 4 to
"keep intact all notices".
c) You must license the entire work, as a whole, under this
License to anyone who comes into possession of a copy. This
License will therefore apply, along with any applicable section 7
additional terms, to the whole of the work, and all its parts,
regardless of how they are packaged. This License gives no
permission to license the work in any other way, but it does not
invalidate such permission if you have separately received it.
d) If the work has interactive user interfaces, each must display
Appropriate Legal Notices; however, if the Program has interactive
interfaces that do not display Appropriate Legal Notices, your
work need not make them do so.
A compilation of a covered work with other separate and independent
works, which are not by their nature extensions of the covered work,
and which are not combined with it such as to form a larger program,
in or on a volume of a storage or distribution medium, is called an
"aggregate" if the compilation and its resulting copyright are not
used to limit the access or legal rights of the compilation's users
beyond what the individual works permit. Inclusion of a covered work
in an aggregate does not cause this License to apply to the other
parts of the aggregate.
6. Conveying Non-Source Forms.
You may convey a covered work in object code form under the terms
of sections 4 and 5, provided that you also convey the
machine-readable Corresponding Source under the terms of this License,
in one of these ways:
a) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by the
Corresponding Source fixed on a durable physical medium
customarily used for software interchange.
b) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by a
written offer, valid for at least three years and valid for as
long as you offer spare parts or customer support for that product
model, to give anyone who possesses the object code either (1) a
copy of the Corresponding Source for all the software in the
product that is covered by this License, on a durable physical
medium customarily used for software interchange, for a price no
more than your reasonable cost of physically performing this
conveying of source, or (2) access to copy the
Corresponding Source from a network server at no charge.
c) Convey individual copies of the object code with a copy of the
written offer to provide the Corresponding Source. This
alternative is allowed only occasionally and noncommercially, and
only if you received the object code with such an offer, in accord
with subsection 6b.
d) Convey the object code by offering access from a designated
place (gratis or for a charge), and offer equivalent access to the
Corresponding Source in the same way through the same place at no
further charge. You need not require recipients to copy the
Corresponding Source along with the object code. If the place to
copy the object code is a network server, the Corresponding Source
may be on a different server (operated by you or a third party)
that supports equivalent copying facilities, provided you maintain
clear directions next to the object code saying where to find the
Corresponding Source. Regardless of what server hosts the
Corresponding Source, you remain obligated to ensure that it is
available for as long as needed to satisfy these requirements.
e) Convey the object code using peer-to-peer transmission, provided
you inform other peers where the object code and Corresponding
Source of the work are being offered to the general public at no
charge under subsection 6d.
A separable portion of the object code, whose source code is excluded
from the Corresponding Source as a System Library, need not be
included in conveying the object code work.
A "User Product" is either (1) a "consumer product", which means any
tangible personal property which is normally used for personal, family,
or household purposes, or (2) anything designed or sold for incorporation
into a dwelling. In determining whether a product is a consumer product,
doubtful cases shall be resolved in favor of coverage. For a particular
product received by a particular user, "normally used" refers to a
typical or common use of that class of product, regardless of the status
of the particular user or of the way in which the particular user
actually uses, or expects or is expected to use, the product. A product
is a consumer product regardless of whether the product has substantial
commercial, industrial or non-consumer uses, unless such uses represent
the only significant mode of use of the product.
"Installation Information" for a User Product means any methods,
procedures, authorization keys, or other information required to install
and execute modified versions of a covered work in that User Product from
a modified version of its Corresponding Source. The information must
suffice to ensure that the continued functioning of the modified object
code is in no case prevented or interfered with solely because
modification has been made.
If you convey an object code work under this section in, or with, or
specifically for use in, a User Product, and the conveying occurs as
part of a transaction in which the right of possession and use of the
User Product is transferred to the recipient in perpetuity or for a
fixed term (regardless of how the transaction is characterized), the
Corresponding Source conveyed under this section must be accompanied
by the Installation Information. But this requirement does not apply
if neither you nor any third party retains the ability to install
modified object code on the User Product (for example, the work has
been installed in ROM).
The requirement to provide Installation Information does not include a
requirement to continue to provide support service, warranty, or updates
for a work that has been modified or installed by the recipient, or for
the User Product in which it has been modified or installed. Access to a
network may be denied when the modification itself materially and
adversely affects the operation of the network or violates the rules and
protocols for communication across the network.
Corresponding Source conveyed, and Installation Information provided,
in accord with this section must be in a format that is publicly
documented (and with an implementation available to the public in
source code form), and must require no special password or key for
unpacking, reading or copying.
7. Additional Terms.
"Additional permissions" are terms that supplement the terms of this
License by making exceptions from one or more of its conditions.
Additional permissions that are applicable to the entire Program shall
be treated as though they were included in this License, to the extent
that they are valid under applicable law. If additional permissions
apply only to part of the Program, that part may be used separately
under those permissions, but the entire Program remains governed by
this License without regard to the additional permissions.
When you convey a copy of a covered work, you may at your option
remove any additional permissions from that copy, or from any part of
it. (Additional permissions may be written to require their own
removal in certain cases when you modify the work.) You may place
additional permissions on material, added by you to a covered work,
for which you have or can give appropriate copyright permission.
Notwithstanding any other provision of this License, for material you
add to a covered work, you may (if authorized by the copyright holders of
that material) supplement the terms of this License with terms:
a) Disclaiming warranty or limiting liability differently from the
terms of sections 15 and 16 of this License; or
b) Requiring preservation of specified reasonable legal notices or
author attributions in that material or in the Appropriate Legal
Notices displayed by works containing it; or
c) Prohibiting misrepresentation of the origin of that material, or
requiring that modified versions of such material be marked in
reasonable ways as different from the original version; or
d) Limiting the use for publicity purposes of names of licensors or
authors of the material; or
e) Declining to grant rights under trademark law for use of some
trade names, trademarks, or service marks; or
f) Requiring indemnification of licensors and authors of that
material by anyone who conveys the material (or modified versions of
it) with contractual assumptions of liability to the recipient, for
any liability that these contractual assumptions directly impose on
those licensors and authors.
All other non-permissive additional terms are considered "further
restrictions" within the meaning of section 10. If the Program as you
received it, or any part of it, contains a notice stating that it is
governed by this License along with a term that is a further
restriction, you may remove that term. If a license document contains
a further restriction but permits relicensing or conveying under this
License, you may add to a covered work material governed by the terms
of that license document, provided that the further restriction does
not survive such relicensing or conveying.
If you add terms to a covered work in accord with this section, you
must place, in the relevant source files, a statement of the
additional terms that apply to those files, or a notice indicating
where to find the applicable terms.
Additional terms, permissive or non-permissive, may be stated in the
form of a separately written license, or stated as exceptions;
the above requirements apply either way.
8. Termination.
You may not propagate or modify a covered work except as expressly
provided under this License. Any attempt otherwise to propagate or
modify it is void, and will automatically terminate your rights under
this License (including any patent licenses granted under the third
paragraph of section 11).
However, if you cease all violation of this License, then your
license from a particular copyright holder is reinstated (a)
provisionally, unless and until the copyright holder explicitly and
finally terminates your license, and (b) permanently, if the copyright
holder fails to notify you of the violation by some reasonable means
prior to 60 days after the cessation.
Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.
Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License. If your rights have been terminated and not permanently
reinstated, you do not qualify to receive new licenses for the same
material under section 10.
9. Acceptance Not Required for Having Copies.
You are not required to accept this License in order to receive or
run a copy of the Program. Ancillary propagation of a covered work
occurring solely as a consequence of using peer-to-peer transmission
to receive a copy likewise does not require acceptance. However,
nothing other than this License grants you permission to propagate or
modify any covered work. These actions infringe copyright if you do
not accept this License. Therefore, by modifying or propagating a
covered work, you indicate your acceptance of this License to do so.
10. Automatic Licensing of Downstream Recipients.
Each time you convey a covered work, the recipient automatically
receives a license from the original licensors, to run, modify and
propagate that work, subject to this License. You are not responsible
for enforcing compliance by third parties with this License.
An "entity transaction" is a transaction transferring control of an
organization, or substantially all assets of one, or subdividing an
organization, or merging organizations. If propagation of a covered
work results from an entity transaction, each party to that
transaction who receives a copy of the work also receives whatever
licenses to the work the party's predecessor in interest had or could
give under the previous paragraph, plus a right to possession of the
Corresponding Source of the work from the predecessor in interest, if
the predecessor has it or can get it with reasonable efforts.
You may not impose any further restrictions on the exercise of the
rights granted or affirmed under this License. For example, you may
not impose a license fee, royalty, or other charge for exercise of
rights granted under this License, and you may not initiate litigation
(including a cross-claim or counterclaim in a lawsuit) alleging that
any patent claim is infringed by making, using, selling, offering for
sale, or importing the Program or any portion of it.
11. Patents.
A "contributor" is a copyright holder who authorizes use under this
License of the Program or a work on which the Program is based. The
work thus licensed is called the contributor's "contributor version".
A contributor's "essential patent claims" are all patent claims
owned or controlled by the contributor, whether already acquired or
hereafter acquired, that would be infringed by some manner, permitted
by this License, of making, using, or selling its contributor version,
but do not include claims that would be infringed only as a
consequence of further modification of the contributor version. For
purposes of this definition, "control" includes the right to grant
patent sublicenses in a manner consistent with the requirements of
this License.
Each contributor grants you a non-exclusive, worldwide, royalty-free
patent license under the contributor's essential patent claims, to
make, use, sell, offer for sale, import and otherwise run, modify and
propagate the contents of its contributor version.
In the following three paragraphs, a "patent license" is any express
agreement or commitment, however denominated, not to enforce a patent
(such as an express permission to practice a patent or covenant not to
sue for patent infringement). To "grant" such a patent license to a
party means to make such an agreement or commitment not to enforce a
patent against the party.
If you convey a covered work, knowingly relying on a patent license,
and the Corresponding Source of the work is not available for anyone
to copy, free of charge and under the terms of this License, through a
publicly available network server or other readily accessible means,
then you must either (1) cause the Corresponding Source to be so
available, or (2) arrange to deprive yourself of the benefit of the
patent license for this particular work, or (3) arrange, in a manner
consistent with the requirements of this License, to extend the patent
license to downstream recipients. "Knowingly relying" means you have
actual knowledge that, but for the patent license, your conveying the
covered work in a country, or your recipient's use of the covered work
in a country, would infringe one or more identifiable patents in that
country that you have reason to believe are valid.
If, pursuant to or in connection with a single transaction or
arrangement, you convey, or propagate by procuring conveyance of, a
covered work, and grant a patent license to some of the parties
receiving the covered work authorizing them to use, propagate, modify
or convey a specific copy of the covered work, then the patent license
you grant is automatically extended to all recipients of the covered
work and works based on it.
A patent license is "discriminatory" if it does not include within
the scope of its coverage, prohibits the exercise of, or is
conditioned on the non-exercise of one or more of the rights that are
specifically granted under this License. You may not convey a covered
work if you are a party to an arrangement with a third party that is
in the business of distributing software, under which you make payment
to the third party based on the extent of your activity of conveying
the work, and under which the third party grants, to any of the
parties who would receive the covered work from you, a discriminatory
patent license (a) in connection with copies of the covered work
conveyed by you (or copies made from those copies), or (b) primarily
for and in connection with specific products or compilations that
contain the covered work, unless you entered into that arrangement,
or that patent license was granted, prior to 28 March 2007.
Nothing in this License shall be construed as excluding or limiting
any implied license or other defenses to infringement that may
otherwise be available to you under applicable patent law.
12. No Surrender of Others' Freedom.
If conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot convey a
covered work so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you may
not convey it at all. For example, if you agree to terms that obligate you
to collect a royalty for further conveying from those to whom you convey
the Program, the only way you could satisfy both those terms and this
License would be to refrain entirely from conveying the Program.
13. Remote Network Interaction; Use with the GNU General Public License.
Notwithstanding any other provision of this License, if you modify the
Program, your modified version must prominently offer all users
interacting with it remotely through a computer network (if your version
supports such interaction) an opportunity to receive the Corresponding
Source of your version by providing access to the Corresponding Source
from a network server at no charge, through some standard or customary
means of facilitating copying of software. This Corresponding Source
shall include the Corresponding Source for any work covered by version 3
of the GNU General Public License that is incorporated pursuant to the
following paragraph.
Notwithstanding any other provision of this License, you have
permission to link or combine any covered work with a work licensed
under version 3 of the GNU General Public License into a single
combined work, and to convey the resulting work. The terms of this
License will continue to apply to the part which is the covered work,
but the work with which it is combined will remain governed by version
3 of the GNU General Public License.
14. Revised Versions of this License.
The Free Software Foundation may publish revised and/or new versions of
the GNU Affero General Public License from time to time. Such new versions
will be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.
Each version is given a distinguishing version number. If the
Program specifies that a certain numbered version of the GNU Affero General
Public License "or any later version" applies to it, you have the
option of following the terms and conditions either of that numbered
version or of any later version published by the Free Software
Foundation. If the Program does not specify a version number of the
GNU Affero General Public License, you may choose any version ever published
by the Free Software Foundation.
If the Program specifies that a proxy can decide which future
versions of the GNU Affero General Public License can be used, that proxy's
public statement of acceptance of a version permanently authorizes you
to choose that version for the Program.
Later license versions may give you additional or different
permissions. However, no additional obligations are imposed on any
author or copyright holder as a result of your choosing to follow a
later version.
15. Disclaimer of Warranty.
THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
16. Limitation of Liability.
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
SUCH DAMAGES.
17. Interpretation of Sections 15 and 16.
If the disclaimer of warranty and limitation of liability provided
above cannot be given local legal effect according to their terms,
reviewing courts shall apply local law that most closely approximates
an absolute waiver of all civil liability in connection with the
Program, unless a warranty or assumption of liability accompanies a
copy of the Program in return for a fee.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest
to attach them to the start of each source file to most effectively
state the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) <year> <name of author>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published
by the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <https://www.gnu.org/licenses/>.
Also add information on how to contact you by electronic and paper mail.
If your software can interact with users remotely through a computer
network, you should also make sure that it provides a way for users to
get its source. For example, if your program is a web application, its
interface could display a "Source" link that leads users to an archive
of the code. There are many ways you could offer source, and different
solutions will be better for different programs; see section 13 for the
specific requirements.
You should also get your employer (if you work as a programmer) or school,
if any, to sign a "copyright disclaimer" for the program, if necessary.
For more information on this, and how to apply and follow the GNU AGPL, see
<https://www.gnu.org/licenses/>.

39
PROJECTS/advanced/ai-threat-detection/README.md Normal file
View File

@ -0,0 +1,39 @@
# AngelusVigil
IN PROGRESS
AI-powered threat detection engine that analyzes web server access logs using machine learning to classify HTTP traffic as benign or malicious in real-time.
Deploys as a Docker sidecar alongside any nginx-based infrastructure. Zero code changes to the monitored application.
## Tech Stack
| Layer | Technology |
|-------|-----------|
| API | FastAPI (async) |
| ML | PyTorch autoencoder + scikit-learn (RF + IF) |
| Inference | ONNX Runtime (CPU) |
| Database | PostgreSQL 18 |
| Cache | Redis 7.4 |
| GeoIP | MaxMind GeoLite2 |
## Quick Start
```bash
just setup
just dev-up
```
## Architecture
3-model ensemble (autoencoder + Random Forest + Isolation Forest) scores each request through a weighted fusion producing a unified threat score [0.0, 1.0]:
- **HIGH** (0.7+): Store + alert + block recommendation
- **MEDIUM** (0.5-0.7): Store + monitor
- **LOW** (<0.5): Log only
See `learn/` for detailed documentation.
## License
AGPLv3 - See [LICENSE](LICENSE)

39
PROJECTS/advanced/ai-threat-detection/backend/alembic.ini Normal file
View File

@ -0,0 +1,39 @@
# ©AngelaMos | 2026
# alembic.ini
[alembic]
script_location = alembic
sqlalchemy.url = postgresql+asyncpg://vigil:changeme@localhost:5432/angelusvigil
[loggers]
keys = root,sqlalchemy,alembic
[handlers]
keys = console
[formatters]
keys = generic
[logger_root]
level = WARN
handlers = console
[logger_sqlalchemy]
level = WARN
handlers =
qualname = sqlalchemy.engine
[logger_alembic]
level = INFO
handlers =
qualname = alembic
[handler_console]
class = StreamHandler
args = (sys.stderr,)
level = NOTSET
formatter = generic
[formatter_generic]
format = %(levelname)-5.5s [%(name)s] %(message)s
datefmt = %H:%M:%S

70
PROJECTS/advanced/ai-threat-detection/backend/alembic/env.py Normal file
View File

@ -0,0 +1,70 @@
"""
©AngelaMos | 2026
env.py
"""
import asyncio
from logging.config import fileConfig
from alembic import context
from sqlalchemy import pool
from sqlalchemy.ext.asyncio import create_async_engine
from sqlmodel import SQLModel
from app.config import settings
from app.models import ModelMetadata, ThreatEvent
_ = (ModelMetadata, ThreatEvent)
target_metadata = SQLModel.metadata
config = context.config
fileConfig(config.config_file_name)
def run_migrations_offline() -> None:
"""
Run migrations in offline mode for SQL script generation.
"""
context.configure(
url=settings.database_url,
target_metadata=target_metadata,
literal_binds=True,
dialect_opts={"paramstyle": "named"},
)
with context.begin_transaction():
context.run_migrations()
def do_run_migrations(connection) -> None:
"""
Execute migrations against a synchronous connection.
"""
context.configure(
connection=connection,
target_metadata=target_metadata,
)
with context.begin_transaction():
context.run_migrations()
async def run_migrations_online() -> None:
"""
Run migrations in online mode using an async engine.
"""
engine = create_async_engine(
settings.database_url,
poolclass=pool.NullPool,
)
async with engine.connect() as connection:
await connection.run_sync(do_run_migrations)
await engine.dispose()
if context.is_offline_mode():
run_migrations_offline()
else:
asyncio.run(run_migrations_online())

28
PROJECTS/advanced/ai-threat-detection/backend/alembic/script.py.mako Normal file
View File

@ -0,0 +1,28 @@
"""
©AngelaMos | 2026
${message}
Revision ID: ${up_revision}
Revises: ${down_revision | comma,n}
Create Date: ${create_date}
"""
from typing import Sequence, Union
from alembic import op
import sqlalchemy as sa
import sqlmodel
${imports if imports else ""}
revision: str = ${repr(up_revision)}
down_revision: Union[str, None] = ${repr(down_revision)}
branch_labels: Union[str, Sequence[str], None] = ${repr(branch_labels)}
depends_on: Union[str, Sequence[str], None] = ${repr(depends_on)}
def upgrade() -> None:
${upgrades if upgrades else "pass"}
def downgrade() -> None:
${downgrades if downgrades else "pass"}

143
PROJECTS/advanced/ai-threat-detection/backend/alembic/versions/65c8ac60f6f6_initial_schema.py Normal file
View File

@ -0,0 +1,143 @@
"""
©AngelaMos | 2026
initial schema
Revision ID: 65c8ac60f6f6
Revises:
Create Date: 2026-02-11 17:43:24.263837
"""
from typing import Sequence, Union
from alembic import op
import sqlalchemy as sa
import sqlmodel
revision: str = "65c8ac60f6f6"
down_revision: Union[str, None] = None
branch_labels: Union[str, Sequence[str], None] = None
depends_on: Union[str, Sequence[str], None] = None
def upgrade() -> None:
op.create_table(
"threat_events",
sa.Column("id", sa.Uuid(), nullable=False),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
server_default=sa.text("CURRENT_TIMESTAMP"),
nullable=False,
),
sa.Column(
"source_ip", sqlmodel.sql.sqltypes.AutoString(length=45), nullable=False
),
sa.Column(
"request_method",
sqlmodel.sql.sqltypes.AutoString(length=10),
nullable=False,
),
sa.Column(
"request_path", sqlmodel.sql.sqltypes.AutoString(), nullable=False
),
sa.Column("status_code", sa.SmallInteger(), nullable=False),
sa.Column("response_size", sa.Integer(), nullable=False),
sa.Column(
"user_agent", sqlmodel.sql.sqltypes.AutoString(), nullable=False
),
sa.Column("threat_score", sa.Float(), nullable=False),
sa.Column(
"severity", sqlmodel.sql.sqltypes.AutoString(length=6), nullable=False
),
sa.Column("component_scores", sa.JSON(), nullable=False),
sa.Column(
"geo_country",
sqlmodel.sql.sqltypes.AutoString(length=2),
nullable=True,
),
sa.Column(
"geo_city",
sqlmodel.sql.sqltypes.AutoString(length=255),
nullable=True,
),
sa.Column("geo_lat", sa.Float(), nullable=True),
sa.Column("geo_lon", sa.Float(), nullable=True),
sa.Column("feature_vector", sa.JSON(), nullable=False),
sa.Column("matched_rules", sa.JSON(), nullable=True),
sa.Column(
"model_version",
sqlmodel.sql.sqltypes.AutoString(length=64),
nullable=True,
),
sa.Column("reviewed", sa.Boolean(), nullable=False),
sa.Column(
"review_label",
sqlmodel.sql.sqltypes.AutoString(length=20),
nullable=True,
),
sa.PrimaryKeyConstraint("id"),
)
op.create_index("idx_threat_events_created_at", "threat_events", ["created_at"])
op.create_index("idx_threat_events_source_ip", "threat_events", ["source_ip"])
op.create_index("idx_threat_events_severity", "threat_events", ["severity"])
op.create_index("idx_threat_events_score", "threat_events", ["threat_score"])
op.create_index(
"idx_threat_events_reviewed",
"threat_events",
["reviewed"],
postgresql_where=sa.text("reviewed = FALSE"),
)
op.create_table(
"model_metadata",
sa.Column("id", sa.Uuid(), nullable=False),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
server_default=sa.text("CURRENT_TIMESTAMP"),
nullable=False,
),
sa.Column(
"model_type",
sqlmodel.sql.sqltypes.AutoString(length=30),
nullable=False,
),
sa.Column(
"version", sqlmodel.sql.sqltypes.AutoString(length=64), nullable=False
),
sa.Column("training_samples", sa.Integer(), nullable=False),
sa.Column("metrics", sa.JSON(), nullable=False),
sa.Column(
"artifact_path", sqlmodel.sql.sqltypes.AutoString(), nullable=False
),
sa.Column("is_active", sa.Boolean(), nullable=False),
sa.Column(
"mlflow_run_id",
sqlmodel.sql.sqltypes.AutoString(length=64),
nullable=True,
),
sa.Column("threshold", sa.Float(), nullable=True),
sa.Column("notes", sqlmodel.sql.sqltypes.AutoString(), nullable=True),
sa.PrimaryKeyConstraint("id"),
)
op.create_index(
"idx_model_metadata_active",
"model_metadata",
["model_type"],
unique=True,
postgresql_where=sa.text("is_active = TRUE"),
)
def downgrade() -> None:
op.drop_index("idx_model_metadata_active", table_name="model_metadata")
op.drop_table("model_metadata")
op.drop_index("idx_threat_events_reviewed", table_name="threat_events")
op.drop_index("idx_threat_events_score", table_name="threat_events")
op.drop_index("idx_threat_events_severity", table_name="threat_events")
op.drop_index("idx_threat_events_source_ip", table_name="threat_events")
op.drop_index("idx_threat_events_created_at", table_name="threat_events")
op.drop_table("threat_events")

14
PROJECTS/advanced/ai-threat-detection/backend/app/__init__.py Normal file
View File

@ -0,0 +1,14 @@
"""
©AngelaMos | 2026
__init__.py
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⣄⠀⠀⠀⣦⣤⣾⣿⠿⠛⣋⣥⣤⣀⠀⠀⠀⠀
⠀⠀⠀⠀⡤⡀⢈⢻⣬⣿⠟⢁⣤⣶⣿⣿⡿⠿⠿⠛⠛⢀⣄⠀
⠀⠀⢢⣘⣿⣿⣶⣿⣯⣤⣾⣿⣿⣿⠟⠁⠄⠀⣾⡇⣼⢻⣿⣾
⣰⠞⠛⢉⣩⣿⣿⣿⣿⣿⣿⣿⣿⠋⣼⣧⣤⣴⠟⣠⣿⢰⣿⣿
⣶⡾⠿⠿⠿⢿⣿⣿⣿⣿⣿⣿⣿⣈⣩⣤⡶⠟⢛⣩⣴⣿⣿⡟
⣠⣄⠈⠀⣰⡦⠙⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣟⡛⠛⠛⠁
⣉⠛⠛⠛⣁⡔⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠥⠀⠀
⣭⣏⣭⣭⣥⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⢠
"""

24
PROJECTS/advanced/ai-threat-detection/backend/app/__main__.py Normal file
View File

@ -0,0 +1,24 @@
"""
©AngelaMos | 2026
__main__.py
"""
import uvicorn
from app.config import settings
def main() -> None:
"""
Run the AngelusVigil API server
"""
uvicorn.run(
"app.main:app",
host=settings.host,
port=settings.port,
reload=settings.debug,
)
if __name__ == "__main__":
main()

4
PROJECTS/advanced/ai-threat-detection/backend/app/api/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

72
PROJECTS/advanced/ai-threat-detection/backend/app/api/health.py Normal file
View File

@ -0,0 +1,72 @@
"""
©AngelaMos | 2026
health.py
"""
import time
from fastapi import APIRouter, Request, Response
from app.core.redis_manager import redis_manager
router = APIRouter()
@router.get("/health")
async def health(request: Request) -> dict:
"""
Liveness probe — returns 200 if the process is alive.
"""
uptime = time.monotonic() - request.app.state.startup_time
return {
"status": "healthy",
"uptime_seconds": round(uptime, 2),
"pipeline_running": request.app.state.pipeline_running,
}
@router.get("/ready")
async def ready(request: Request, response: Response) -> dict:
"""
Readiness probe — checks all service dependencies.
"""
redis_ok = await _check_redis()
database_ok = await _check_database(request)
checks = {
"database": "ok" if database_ok else "error",
"redis": "ok" if redis_ok else "error",
"models_loaded": getattr(request.app.state, "models_loaded", False),
}
all_ok = database_ok and redis_ok
if not all_ok:
response.status_code = 503
return {
"status": "ready" if all_ok else "not_ready",
"checks": checks,
}
async def _check_redis() -> bool:
"""
Ping Redis and return connectivity status.
"""
return await redis_manager.ping()
async def _check_database(request: Request) -> bool:
"""
Verify database engine is connected and responsive.
"""
engine = getattr(request.app.state, "db_engine", None)
if engine is None:
return False
try:
async with engine.connect() as conn:
await conn.execute("SELECT 1")
return True
except Exception:
return False

47
PROJECTS/advanced/ai-threat-detection/backend/app/config.py Normal file
View File

@ -0,0 +1,47 @@
"""
©AngelaMos | 2026
config.py
"""
from pydantic_settings import BaseSettings, SettingsConfigDict
class Settings(BaseSettings):
"""
Application configuration loaded from environment variables.
"""
model_config = SettingsConfigDict(
env_file=".env",
env_file_encoding="utf-8",
case_sensitive=False,
)
app_name: str = "AngelusVigil"
env: str = "development"
debug: bool = False
host: str = "0.0.0.0"
port: int = 8000
api_key: str = ""
log_level: str = "INFO"
database_url: str = (
"postgresql+asyncpg://vigil:changeme@localhost:5432/angelusvigil"
)
redis_url: str = "redis://localhost:6379"
geoip_db_path: str = "/usr/share/GeoIP/GeoLite2-City.mmdb"
nginx_log_path: str = "/var/log/nginx/access.log"
raw_queue_size: int = 1000
parsed_queue_size: int = 500
feature_queue_size: int = 200
alert_queue_size: int = 100
batch_size: int = 32
batch_timeout_ms: int = 50
settings = Settings()

4
PROJECTS/advanced/ai-threat-detection/backend/app/core/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

4
PROJECTS/advanced/ai-threat-detection/backend/app/core/detection/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

131
PROJECTS/advanced/ai-threat-detection/backend/app/core/detection/rules.py Normal file
View File

@ -0,0 +1,131 @@
"""
©AngelaMos | 2026
rules.py
"""
from dataclasses import dataclass, field
from typing import NamedTuple
from app.core.features.patterns import (
COMMAND_INJECTION,
DOUBLE_ENCODED,
FILE_INCLUSION,
PATH_TRAVERSAL,
SQLI,
XSS,
)
from app.core.features.signatures import SCANNER_USER_AGENTS
from app.core.ingestion.parsers import ParsedLogEntry
class _PatternRule(NamedTuple):
"""
A regex-based detection rule applied to the request URI.
"""
name: str
pattern: object
score: float
class _ThresholdRule(NamedTuple):
"""
A threshold-based detection rule applied to a windowed feature.
"""
name: str
feature_key: str
threshold: float
score: float
_PATTERN_RULES: list[_PatternRule] = [
_PatternRule("COMMAND_INJECTION", COMMAND_INJECTION, 0.90),
_PatternRule("SQL_INJECTION", SQLI, 0.85),
_PatternRule("XSS", XSS, 0.80),
_PatternRule("FILE_INCLUSION", FILE_INCLUSION, 0.75),
_PatternRule("PATH_TRAVERSAL", PATH_TRAVERSAL, 0.60),
]
_THRESHOLD_RULES: list[_ThresholdRule] = [
_ThresholdRule("RATE_ANOMALY", "req_count_1m", 100.0, 0.30),
_ThresholdRule("HIGH_ERROR_RATE", "error_rate_5m", 0.5, 0.25),
]
_DOUBLE_ENCODING_SCORE = 0.40
_SCANNER_UA_SCORE = 0.35
_BOOST_PER_ADDITIONAL_RULE = 0.05
@dataclass(frozen=True, slots=True)
class RuleResult:
"""
Output of the rule-based detection engine for a single request.
"""
threat_score: float
severity: str
matched_rules: list[str] = field(default_factory=list)
component_scores: dict[str, float] = field(default_factory=dict)
def _classify_severity(score: float) -> str:
"""
Map a threat score to a severity label.
"""
if score >= 0.7:
return "HIGH"
if score >= 0.5:
return "MEDIUM"
return "LOW"
class RuleEngine:
"""
Cold-start rule-based detection engine inspired by ModSecurity CRS.
Scores requests using pattern matching, signature detection,
and behavioral thresholds from windowed features.
"""
def score_request(
self,
features: dict[str, int | float | bool | str],
entry: ParsedLogEntry,
) -> RuleResult:
"""
Evaluate all rules against a request and return a composite score.
"""
matched: list[tuple[str, float]] = []
uri = entry.path
if entry.query_string:
uri = f"{entry.path}?{entry.query_string}"
for rule in _PATTERN_RULES:
if rule.pattern.search(uri):
matched.append((rule.name, rule.score))
if DOUBLE_ENCODED.search(uri):
matched.append(("DOUBLE_ENCODING", _DOUBLE_ENCODING_SCORE))
ua_lower = entry.user_agent.lower()
if any(sig in ua_lower for sig in SCANNER_USER_AGENTS):
matched.append(("SCANNER_UA", _SCANNER_UA_SCORE))
for rule in _THRESHOLD_RULES:
value = features.get(rule.feature_key, 0)
if isinstance(value, (int, float)) and value > rule.threshold:
matched.append((rule.name, rule.score))
if not matched:
return RuleResult(threat_score=0.0, severity="LOW")
scores = sorted([s for _, s in matched], reverse=True)
threat_score = min(
scores[0] + _BOOST_PER_ADDITIONAL_RULE * (len(scores) - 1),
1.0,
)
return RuleResult(
threat_score=threat_score,
severity=_classify_severity(threat_score),
matched_rules=[name for name, _ in matched],
component_scores={name: score for name, score in matched},
)

4
PROJECTS/advanced/ai-threat-detection/backend/app/core/enrichment/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

88
PROJECTS/advanced/ai-threat-detection/backend/app/core/enrichment/geoip.py Normal file
View File

@ -0,0 +1,88 @@
"""
©AngelaMos | 2026
geoip.py
"""
import asyncio
import ipaddress
import logging
from dataclasses import dataclass
from pathlib import Path
import geoip2.database
from geoip2.errors import AddressNotFoundError
logger = logging.getLogger(__name__)
@dataclass(frozen=True, slots=True)
class GeoResult:
"""
Structured GeoIP lookup result.
"""
country: str | None
city: str | None
lat: float | None
lon: float | None
class GeoIPService:
"""
Async GeoIP lookup service backed by a local MaxMind GeoLite2-City database.
"""
def __init__(self, db_path: str) -> None:
self._reader: geoip2.database.Reader | None = None
self._db_path = db_path
if Path(db_path).is_file():
self._reader = geoip2.database.Reader(db_path)
logger.info("GeoIP database loaded from %s", db_path)
else:
logger.warning("GeoIP database not found at %s — lookups disabled", db_path)
async def lookup(self, ip: str) -> GeoResult | None:
"""
Look up geographic data for an IP address.
Returns None for private IPs, unknown addresses, or when no database is loaded.
"""
if self._reader is None:
return None
try:
if ipaddress.ip_address(ip).is_private:
return None
except ValueError:
return None
try:
response = await asyncio.to_thread(self._reader.city, ip)
except AddressNotFoundError:
return None
return GeoResult(
country=response.country.iso_code,
city=response.city.name,
lat=response.location.latitude,
lon=response.location.longitude,
)
def swap_reader(self, new_path: str) -> None:
"""
Atomically replace the database reader after a .mmdb file update.
"""
old_reader = self._reader
self._reader = geoip2.database.Reader(new_path)
self._db_path = new_path
logger.info("GeoIP reader swapped to %s", new_path)
if old_reader is not None:
old_reader.close()
def close(self) -> None:
"""
Close the underlying database reader.
"""
if self._reader is not None:
self._reader.close()
self._reader = None

4
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

206
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/aggregator.py Normal file
View File

@ -0,0 +1,206 @@
"""
©AngelaMos | 2026
aggregator.py
"""
import hashlib
import math
from collections import Counter
import redis.asyncio as aioredis
WINDOW_1M = 60
WINDOW_5M = 300
WINDOW_10M = 600
KEY_TTL = 900
def _hash_member(value: str) -> str:
"""
Produce a compact 16-char hex digest for sorted set deduplication.
"""
return hashlib.md5(value.encode()).hexdigest()[:16]
class WindowAggregator:
"""
Per-IP sliding window feature aggregator backed by Redis sorted sets.
"""
def __init__(self, redis_client: aioredis.Redis) -> None:
self._redis = redis_client
async def record_and_aggregate(
self,
ip: str,
request_id: str,
path: str,
path_depth: int,
method: str,
status_code: int,
user_agent: str,
response_size: int,
timestamp: float,
) -> dict[str, float]:
"""
Record a request into Redis sorted sets and return all 12
per-IP windowed features.
"""
prefix = f"ip:{ip}"
keys = {
"requests": f"{prefix}:requests",
"paths": f"{prefix}:paths",
"statuses": f"{prefix}:statuses",
"uas": f"{prefix}:uas",
"sizes": f"{prefix}:sizes",
"methods": f"{prefix}:methods",
"depths": f"{prefix}:depths",
}
trim_boundary = timestamp - KEY_TTL
w1m = timestamp - WINDOW_1M
w5m = timestamp - WINDOW_5M
w10m = timestamp - WINDOW_10M
pipe = self._redis.pipeline()
pipe.zadd(keys["requests"], {request_id: timestamp})
pipe.zadd(keys["paths"], {_hash_member(path): timestamp})
pipe.zadd(
keys["statuses"], {f"{status_code}:{request_id}": timestamp}
)
pipe.zadd(keys["uas"], {_hash_member(user_agent): timestamp})
pipe.zadd(
keys["sizes"], {f"{response_size}:{request_id}": timestamp}
)
pipe.zadd(keys["methods"], {f"{method}:{request_id}": timestamp})
pipe.zadd(
keys["depths"], {f"{path_depth}:{request_id}": timestamp}
)
for key in keys.values():
pipe.zremrangebyscore(key, "-inf", trim_boundary)
pipe.zcount(keys["requests"], w1m, "+inf")
pipe.zcount(keys["requests"], w5m, "+inf")
pipe.zcount(keys["requests"], w10m, "+inf")
pipe.zcount(keys["paths"], w5m, "+inf")
pipe.zcount(keys["uas"], w10m, "+inf")
pipe.zrangebyscore(keys["statuses"], w5m, "+inf")
pipe.zrangebyscore(keys["sizes"], w5m, "+inf")
pipe.zrangebyscore(keys["methods"], w5m, "+inf")
pipe.zrangebyscore(keys["depths"], w5m, "+inf")
pipe.zrangebyscore(
keys["requests"], w10m, "+inf", withscores=True
)
for key in keys.values():
pipe.expire(key, KEY_TTL)
results = await pipe.execute()
read_start = 14
req_count_1m = results[read_start]
req_count_5m = results[read_start + 1]
req_count_10m = results[read_start + 2]
unique_paths_5m = results[read_start + 3]
unique_uas_10m = results[read_start + 4]
statuses_5m = results[read_start + 5]
sizes_5m = results[read_start + 6]
methods_5m = results[read_start + 7]
depths_5m = results[read_start + 8]
requests_with_scores = results[read_start + 9]
irt_mean, irt_std = _inter_request_time_stats(requests_with_scores)
return {
"req_count_1m": float(req_count_1m),
"req_count_5m": float(req_count_5m),
"req_count_10m": float(req_count_10m),
"error_rate_5m": _error_rate(statuses_5m),
"unique_paths_5m": float(unique_paths_5m),
"unique_uas_10m": float(unique_uas_10m),
"method_entropy_5m": _method_entropy(methods_5m),
"avg_response_size_5m": _avg_response_size(sizes_5m),
"status_diversity_5m": _status_diversity(statuses_5m),
"path_depth_variance_5m": _path_depth_variance(depths_5m),
"inter_request_time_mean": irt_mean,
"inter_request_time_std": irt_std,
}
def _error_rate(status_members: list[str]) -> float:
"""
Ratio of 4xx/5xx responses to total responses.
"""
if not status_members:
return 0.0
errors = sum(
1 for m in status_members if int(m.split(":")[0]) >= 400
)
return errors / len(status_members)
def _avg_response_size(size_members: list[str]) -> float:
"""
Mean response body size from size:request_id members.
"""
if not size_members:
return 0.0
sizes = [int(m.split(":")[0]) for m in size_members]
return sum(sizes) / len(sizes)
def _method_entropy(method_members: list[str]) -> float:
"""
Shannon entropy of HTTP method distribution.
"""
if not method_members:
return 0.0
methods = [m.split(":")[0] for m in method_members]
counts = Counter(methods)
total = len(methods)
return -sum(
(c / total) * math.log2(c / total) for c in counts.values()
)
def _status_diversity(status_members: list[str]) -> float:
"""
Count of distinct HTTP status codes.
"""
if not status_members:
return 0.0
codes = {m.split(":")[0] for m in status_members}
return float(len(codes))
def _path_depth_variance(depth_members: list[str]) -> float:
"""
Population variance of path depth values.
"""
if len(depth_members) < 2:
return 0.0
depths = [int(m.split(":")[0]) for m in depth_members]
mean = sum(depths) / len(depths)
return sum((d - mean) ** 2 for d in depths) / len(depths)
def _inter_request_time_stats(
entries: list[tuple[str, float]],
) -> tuple[float, float]:
"""
Mean and standard deviation of inter-request intervals in milliseconds.
"""
if len(entries) < 2:
return 0.0, 0.0
timestamps = sorted(score for _, score in entries)
deltas = [
(timestamps[i + 1] - timestamps[i]) * 1000
for i in range(len(timestamps) - 1)
]
mean = sum(deltas) / len(deltas)
if len(deltas) < 2:
return mean, 0.0
variance = sum((d - mean) ** 2 for d in deltas) / len(deltas)
return mean, math.sqrt(variance)

43
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/encoder.py Normal file
View File

@ -0,0 +1,43 @@
"""
©AngelaMos | 2026
encoder.py
"""
from app.core.features.mappings import (
BOOLEAN_FEATURES,
CATEGORICAL_ENCODERS,
FEATURE_ORDER,
)
def _encode_country(code: str) -> float:
"""
Deterministic ordinal encoding for 2-letter ISO country codes.
"""
if not code or len(code) != 2:
return 0.0
return float((ord(code[0]) - 64) * 26 + (ord(code[1]) - 64))
def encode_for_inference(
features: dict[str, int | float | bool | str],
) -> list[float]:
"""
Encode a combined feature dict into a 35-element float vector
matching the model input specification.
"""
vector: list[float] = []
for name in FEATURE_ORDER:
raw = features[name]
if name in BOOLEAN_FEATURES:
vector.append(1.0 if raw else 0.0)
elif name in CATEGORICAL_ENCODERS:
vector.append(float(CATEGORICAL_ENCODERS[name].get(raw, 0)))
elif name == "country_code":
vector.append(_encode_country(raw))
else:
vector.append(float(raw))
return vector

88
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/extractor.py Normal file
View File

@ -0,0 +1,88 @@
"""
©AngelaMos | 2026
extractor.py
"""
import ipaddress
import math
from collections import Counter
from posixpath import splitext
from app.core.features.patterns import (
ATTACK_COMBINED,
DOUBLE_ENCODED,
ENCODED_CHARS,
)
from app.core.features.signatures import BOT_USER_AGENTS, SCANNER_USER_AGENTS
from app.core.ingestion.parsers import ParsedLogEntry
def _shannon_entropy(s: str) -> float:
"""
Compute Shannon entropy of a string.
"""
if not s:
return 0.0
length = len(s)
counts = Counter(s)
return -sum(
(c / length) * math.log2(c / length) for c in counts.values()
)
def _is_private_ip(ip_str: str) -> bool:
"""
Check whether an IP address is in a private or loopback range.
"""
try:
return ipaddress.ip_address(ip_str).is_private
except ValueError:
return False
def extract_request_features(
entry: ParsedLogEntry,
country_code: str = "",
) -> dict[str, int | float | bool | str]:
"""
Extract 23 stateless per-request features from a parsed log entry.
"""
full_uri = entry.path
if entry.query_string:
full_uri = f"{entry.path}?{entry.query_string}"
ua_lower = entry.user_agent.lower()
non_alnum = sum(1 for c in entry.path if not c.isalnum())
path_len = len(entry.path)
_, ext = splitext(entry.path)
return {
"http_method": entry.method,
"path_depth": len([s for s in entry.path.split("/") if s]),
"path_entropy": _shannon_entropy(entry.path),
"path_length": path_len,
"query_string_length": len(entry.query_string),
"query_param_count": (
len(entry.query_string.split("&")) if entry.query_string else 0
),
"has_encoded_chars": bool(ENCODED_CHARS.search(full_uri)),
"has_double_encoding": bool(DOUBLE_ENCODED.search(full_uri)),
"status_code": entry.status_code,
"status_class": f"{entry.status_code // 100}xx",
"response_size": entry.response_size,
"hour_of_day": entry.timestamp.hour,
"day_of_week": entry.timestamp.weekday(),
"is_weekend": entry.timestamp.weekday() >= 5,
"ua_length": len(entry.user_agent),
"ua_entropy": _shannon_entropy(entry.user_agent),
"is_known_bot": any(sig in ua_lower for sig in BOT_USER_AGENTS),
"is_known_scanner": any(
sig in ua_lower for sig in SCANNER_USER_AGENTS
),
"has_attack_pattern": bool(ATTACK_COMBINED.search(full_uri)),
"special_char_ratio": non_alnum / path_len if path_len else 0.0,
"file_extension": ext,
"country_code": country_code,
"is_private_ip": _is_private_ip(entry.ip),
}

104
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/mappings.py Normal file
View File

@ -0,0 +1,104 @@
"""
©AngelaMos | 2026
mappings.py
"""
METHOD_MAP: dict[str, int] = {
"GET": 1,
"POST": 2,
"PUT": 3,
"DELETE": 4,
"PATCH": 5,
"HEAD": 6,
"OPTIONS": 7,
}
STATUS_CLASS_MAP: dict[str, int] = {
"1xx": 1,
"2xx": 2,
"3xx": 3,
"4xx": 4,
"5xx": 5,
}
EXTENSION_MAP: dict[str, int] = {
".html": 1,
".htm": 2,
".php": 3,
".js": 4,
".css": 5,
".json": 6,
".xml": 7,
".jpg": 8,
".jpeg": 9,
".png": 10,
".gif": 11,
".svg": 12,
".ico": 13,
".pdf": 14,
".zip": 15,
".txt": 16,
".asp": 17,
".aspx": 18,
".jsp": 19,
".py": 20,
".rb": 21,
".woff2": 22,
".woff": 23,
".ttf": 24,
".map": 25,
}
FEATURE_ORDER: list[str] = [
"http_method",
"path_depth",
"path_entropy",
"path_length",
"query_string_length",
"query_param_count",
"has_encoded_chars",
"has_double_encoding",
"status_code",
"status_class",
"response_size",
"hour_of_day",
"day_of_week",
"is_weekend",
"ua_length",
"ua_entropy",
"is_known_bot",
"is_known_scanner",
"has_attack_pattern",
"special_char_ratio",
"file_extension",
"country_code",
"is_private_ip",
"req_count_1m",
"req_count_5m",
"req_count_10m",
"error_rate_5m",
"unique_paths_5m",
"unique_uas_10m",
"method_entropy_5m",
"avg_response_size_5m",
"status_diversity_5m",
"path_depth_variance_5m",
"inter_request_time_mean",
"inter_request_time_std",
]
CATEGORICAL_ENCODERS: dict[str, dict[str, int]] = {
"http_method": METHOD_MAP,
"status_class": STATUS_CLASS_MAP,
"file_extension": EXTENSION_MAP,
}
BOOLEAN_FEATURES: frozenset[str] = frozenset({
"has_encoded_chars",
"has_double_encoding",
"is_weekend",
"is_known_bot",
"is_known_scanner",
"has_attack_pattern",
"is_private_ip",
})

96
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/patterns.py Normal file
View File

@ -0,0 +1,96 @@
"""
©AngelaMos | 2026
patterns.py
"""
import re
ENCODED_CHARS = re.compile(r"%[0-9a-fA-F]{2}")
DOUBLE_ENCODED = re.compile(r"%25[0-9a-fA-F]{2}")
_SQLI = (
r"(?:union\s+(?:all\s+)?select|"
r"'\s*or\s+.+=|"
r"'\s*and\s+.+=|"
r"1\s*=\s*1|"
r"sleep\s*\(|"
r"benchmark\s*\(|"
r"waitfor\s+delay|"
r"extractvalue\s*\(|"
r"updatexml\s*\(|"
r"load_file\s*\(|"
r"into\s+(?:out|dump)file|"
r"group\s+by\s+.+having|"
r"order\s+by\s+\d+|"
r"(?:drop|alter|create)\s+table|"
r"information_schema|"
r"(?:char|concat|hex|unhex)\s*\(|"
r"0x[0-9a-f]{6,}|"
r"--\s*$|"
r"/\*.*?\*/)"
)
_XSS = (
r"(?:<\s*script|"
r"javascript\s*:|"
r"vbscript\s*:|"
r"on(?:error|load|click|mouse\w+|focus|blur|submit|change)\s*=|"
r"<\s*(?:img|svg|iframe|object|embed|link|style|body|input|form)\b[^>]*\bon\w+\s*=|"
r"<\s*iframe|"
r"<\s*svg\b.*?on\w+\s*=|"
r"document\s*\.\s*(?:cookie|write|location)|"
r"window\s*\.\s*(?:location|open)|"
r"eval\s*\(|"
r"alert\s*\(|"
r"prompt\s*\(|"
r"confirm\s*\(|"
r"expression\s*\(|"
r"String\s*\.\s*fromCharCode)"
)
_PATH_TRAVERSAL = (
r"(?:\.\./|"
r"\.\.\\|"
r"%2e%2e[%/\\]|"
r"%252e%252e|"
r"(?:etc/(?:passwd|shadow|hosts)|"
r"proc/self/|"
r"windows/system32|"
r"boot\.ini|"
r"web\.config|"
r"\.env|"
r"\.git/config|"
r"wp-config\.php))"
)
_COMMAND_INJECTION = (
r"(?:;\s*(?:ls|cat|rm|wget|curl|chmod|chown|nc|bash|sh|python|perl|ruby|php)\b|"
r"\|\s*(?:cat|ls|id|whoami|uname|pwd|env|set|netstat|ifconfig|ip)\b|"
r"\$\(|"
r"`[^`]+`|"
r"\$\{|"
r">\s*/(?:etc|tmp|var)|"
r"&&\s*(?:cat|ls|id|whoami|wget|curl)\b)"
)
_FILE_INCLUSION = (
r"(?:php://|"
r"file://|"
r"data://|"
r"expect://|"
r"input://|"
r"zip://|"
r"phar://|"
r"glob://)"
)
SQLI = re.compile(_SQLI, re.IGNORECASE)
XSS = re.compile(_XSS, re.IGNORECASE)
PATH_TRAVERSAL = re.compile(_PATH_TRAVERSAL, re.IGNORECASE)
COMMAND_INJECTION = re.compile(_COMMAND_INJECTION, re.IGNORECASE)
FILE_INCLUSION = re.compile(_FILE_INCLUSION, re.IGNORECASE)
ATTACK_COMBINED = re.compile(
r"|".join((_SQLI, _XSS, _PATH_TRAVERSAL, _COMMAND_INJECTION, _FILE_INCLUSION)),
re.IGNORECASE,
)

85
PROJECTS/advanced/ai-threat-detection/backend/app/core/features/signatures.py Normal file
View File

@ -0,0 +1,85 @@
"""
©AngelaMos | 2026
signatures.py
"""
BOT_USER_AGENTS: frozenset[str] = frozenset({
"googlebot",
"bingbot",
"slurp",
"duckduckbot",
"baiduspider",
"yandexbot",
"sogou",
"facebot",
"ia_archiver",
"applebot",
"petalbot",
"semrushbot",
"ahrefsbot",
"mj12bot",
"dotbot",
"rogerbot",
"linkedinbot",
"twitterbot",
"gptbot",
"claudebot",
"amazonbot",
"bytespider",
"ccbot",
"dataforseo",
"seznambot",
"megaindex",
"blexbot",
"exabot",
"archive.org_bot",
"mojeekbot",
"uptimerobot",
"deadlinkchecker",
"sitebulb",
"screaming frog",
})
SCANNER_USER_AGENTS: frozenset[str] = frozenset({
"nikto",
"sqlmap",
"nessus",
"openvas",
"acunetix",
"w3af",
"nmap",
"masscan",
"zgrab",
"gobuster",
"dirbuster",
"dirb",
"wfuzz",
"ffuf",
"nuclei",
"burp",
"zap",
"arachni",
"skipfish",
"wpscan",
"joomscan",
"whatweb",
"httprint",
"fierce",
"subfinder",
"amass",
"httpx",
"jaeles",
"xray",
"gau",
"hakrawler",
"katana",
"cariddi",
"gospider",
"feroxbuster",
"rustbuster",
"patator",
"hydra",
"medusa",
"metasploit",
"cobalt",
})

4
PROJECTS/advanced/ai-threat-detection/backend/app/core/ingestion/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

152
PROJECTS/advanced/ai-threat-detection/backend/app/core/ingestion/parsers.py Normal file
View File

@ -0,0 +1,152 @@
"""
©AngelaMos | 2026
parsers.py
"""
import re
from dataclasses import dataclass
from datetime import datetime
@dataclass(frozen=True, slots=True)
class ParsedLogEntry:
"""
Structured representation of a single nginx access log line.
"""
ip: str
timestamp: datetime
method: str
path: str
query_string: str
status_code: int
response_size: int
referer: str
user_agent: str
raw_line: str
_TIMESTAMP_FMT = "%d/%b/%Y:%H:%M:%S %z"
_COMBINED_RE = re.compile(
r"(?P<ip>\S+) \S+ \S+ "
r"\[(?P<timestamp>[^\]]+)\] "
r'"(?P<request>[^"]*)" '
r"(?P<status>\d{3}) "
r"(?P<size>\S+) "
r'"(?P<referer>[^"]*)" '
r'"(?P<user_agent>[^"]*)"'
)
def parse_combined(line: str) -> ParsedLogEntry | None:
"""
Parse an nginx combined-format log line using string-split primary
with compiled regex fallback.
"""
if not line:
return None
result = _parse_split(line)
if result is not None:
return result
return _parse_regex(line)
def _parse_split(line: str) -> ParsedLogEntry | None:
"""
Fast string-split parser for well-formed nginx combined lines.
"""
try:
parts = line.split('"')
if len(parts) < 6:
return None
prefix = parts[0]
request_line = parts[1]
status_size = parts[2]
referer_raw = parts[3]
user_agent = parts[5]
bracket_open = prefix.index("[")
bracket_close = prefix.index("]")
ip = prefix[:bracket_open].split()[0]
timestamp = datetime.strptime(
prefix[bracket_open + 1 : bracket_close], _TIMESTAMP_FMT
)
request_parts = request_line.split(" ", 2)
method = request_parts[0]
full_uri = request_parts[1] if len(request_parts) > 1 else ""
if "?" in full_uri:
path, query_string = full_uri.split("?", 1)
else:
path = full_uri
query_string = ""
tokens = status_size.strip().split()
status_code = int(tokens[0])
response_size = int(tokens[1]) if tokens[1] != "-" else 0
referer = "" if referer_raw == "-" else referer_raw
return ParsedLogEntry(
ip=ip,
timestamp=timestamp,
method=method,
path=path,
query_string=query_string,
status_code=status_code,
response_size=response_size,
referer=referer,
user_agent=user_agent,
raw_line=line,
)
except (ValueError, IndexError):
return None
def _parse_regex(line: str) -> ParsedLogEntry | None:
"""
Regex fallback for non-standard or edge-case log lines.
"""
match = _COMBINED_RE.match(line)
if not match:
return None
try:
timestamp = datetime.strptime(match["timestamp"], _TIMESTAMP_FMT)
request_line = match["request"]
request_parts = request_line.split(" ", 2)
method = request_parts[0]
full_uri = request_parts[1] if len(request_parts) > 1 else ""
if "?" in full_uri:
path, query_string = full_uri.split("?", 1)
else:
path = full_uri
query_string = ""
size_raw = match["size"]
response_size = int(size_raw) if size_raw != "-" else 0
referer_raw = match["referer"]
referer = "" if referer_raw == "-" else referer_raw
return ParsedLogEntry(
ip=match["ip"],
timestamp=timestamp,
method=method,
path=path,
query_string=query_string,
status_code=int(match["status"]),
response_size=response_size,
referer=referer,
user_agent=match["user_agent"],
raw_line=line,
)
except (ValueError, IndexError):
return None

214
PROJECTS/advanced/ai-threat-detection/backend/app/core/ingestion/pipeline.py Normal file
View File

@ -0,0 +1,214 @@
"""
©AngelaMos | 2026
pipeline.py
"""
import asyncio
import logging
import uuid
from collections.abc import Awaitable, Callable
from dataclasses import dataclass
import redis.asyncio as aioredis
from app.core.detection.rules import RuleEngine, RuleResult
from app.core.enrichment.geoip import GeoIPService, GeoResult
from app.core.features.aggregator import WindowAggregator
from app.core.features.encoder import encode_for_inference
from app.core.features.extractor import extract_request_features
from app.core.ingestion.parsers import ParsedLogEntry, parse_combined
logger = logging.getLogger(__name__)
@dataclass(slots=True)
class EnrichedRequest:
"""
A parsed log entry enriched with extracted features and GeoIP data.
"""
entry: ParsedLogEntry
features: dict[str, int | float | bool | str]
feature_vector: list[float]
geo: GeoResult | None
@dataclass(slots=True)
class ScoredRequest:
"""
A fully scored request ready for dispatch.
"""
entry: ParsedLogEntry
features: dict[str, int | float | bool | str]
feature_vector: list[float]
geo: GeoResult | None
rule_result: RuleResult
class Pipeline:
"""
Four-stage async pipeline that transforms raw log lines
into scored threat candidates.
Stages:
raw_queue → [parse] → parsed_queue → [enrich+features]
→ feature_queue → [detect] → alert_queue → [dispatch]
"""
def __init__(
self,
redis_client: aioredis.Redis,
rule_engine: RuleEngine,
geoip: GeoIPService | None = None,
on_result: Callable[[ScoredRequest], Awaitable[None]] | None = None,
raw_queue_size: int = 1000,
parsed_queue_size: int = 500,
feature_queue_size: int = 200,
alert_queue_size: int = 100,
) -> None:
self.raw_queue: asyncio.Queue[str | None] = asyncio.Queue(
maxsize=raw_queue_size,
)
self._parsed_queue: asyncio.Queue[ParsedLogEntry | None] = asyncio.Queue(
maxsize=parsed_queue_size,
)
self._feature_queue: asyncio.Queue[EnrichedRequest | None] = asyncio.Queue(
maxsize=feature_queue_size,
)
self._alert_queue: asyncio.Queue[ScoredRequest | None] = asyncio.Queue(
maxsize=alert_queue_size,
)
self._aggregator = WindowAggregator(redis_client)
self._rule_engine = rule_engine
self._geoip = geoip
self._on_result = on_result
self._tasks: list[asyncio.Task] = []
async def start(self) -> None:
"""
Spawn worker tasks for each pipeline stage.
"""
self._tasks = [
asyncio.create_task(self._parse_worker(), name="parse"),
asyncio.create_task(self._feature_worker(), name="feature"),
asyncio.create_task(self._detection_worker(), name="detection"),
asyncio.create_task(self._dispatch_worker(), name="dispatch"),
]
logger.info("Pipeline started — 4 stage workers running")
async def stop(self) -> None:
"""
Send a poison pill through the chain and wait for all workers to exit.
"""
await self.raw_queue.put(None)
await asyncio.gather(*self._tasks)
logger.info("Pipeline stopped — all workers exited")
async def _parse_worker(self) -> None:
"""
Stage 1: Parse raw log lines into structured entries.
"""
while True:
line = await self.raw_queue.get()
if line is None:
self.raw_queue.task_done()
await self._parsed_queue.put(None)
break
try:
entry = parse_combined(line)
if entry is not None:
await self._parsed_queue.put(entry)
except Exception:
logger.exception("Parse error")
self.raw_queue.task_done()
async def _feature_worker(self) -> None:
"""
Stage 2: Enrich with GeoIP, extract per-request features,
aggregate per-IP windowed features, and encode the 35-dim vector.
"""
while True:
entry = await self._parsed_queue.get()
if entry is None:
self._parsed_queue.task_done()
await self._feature_queue.put(None)
break
try:
country_code = ""
geo = None
if self._geoip is not None:
geo = await self._geoip.lookup(entry.ip)
if geo and geo.country:
country_code = geo.country
per_request = extract_request_features(entry, country_code)
windowed = await self._aggregator.record_and_aggregate(
ip=entry.ip,
request_id=uuid.uuid4().hex,
path=entry.path,
path_depth=per_request["path_depth"],
method=entry.method,
status_code=entry.status_code,
user_agent=entry.user_agent,
response_size=entry.response_size,
timestamp=entry.timestamp.timestamp(),
)
merged = {**per_request, **windowed}
vector = encode_for_inference(merged)
await self._feature_queue.put(
EnrichedRequest(
entry=entry,
features=merged,
feature_vector=vector,
geo=geo,
)
)
except Exception:
logger.exception("Feature extraction failed for %s", entry.ip)
self._parsed_queue.task_done()
async def _detection_worker(self) -> None:
"""
Stage 3: Score enriched requests using the rule engine.
"""
while True:
enriched = await self._feature_queue.get()
if enriched is None:
self._feature_queue.task_done()
await self._alert_queue.put(None)
break
try:
rule_result = self._rule_engine.score_request(
enriched.features, enriched.entry,
)
await self._alert_queue.put(
ScoredRequest(
entry=enriched.entry,
features=enriched.features,
feature_vector=enriched.feature_vector,
geo=enriched.geo,
rule_result=rule_result,
)
)
except Exception:
logger.exception("Detection failed")
self._feature_queue.task_done()
async def _dispatch_worker(self) -> None:
"""
Stage 4: Dispatch scored results via the on_result callback.
"""
while True:
scored = await self._alert_queue.get()
if scored is None:
self._alert_queue.task_done()
break
try:
if self._on_result is not None:
await self._on_result(scored)
except Exception:
logger.exception("Dispatch failed for %s", scored.entry.ip)
self._alert_queue.task_done()

185
PROJECTS/advanced/ai-threat-detection/backend/app/core/ingestion/tailer.py Normal file
View File

@ -0,0 +1,185 @@
"""
©AngelaMos | 2026
tailer.py
"""
import asyncio
import logging
import os
from pathlib import Path
from watchdog.events import (
FileCreatedEvent,
FileModifiedEvent,
FileMovedEvent,
FileSystemEventHandler,
)
from watchdog.observers import Observer
logger = logging.getLogger(__name__)
class _LogHandler(FileSystemEventHandler):
"""
Watchdog event handler that detects modifications and log rotation
for a single target file, pushing new lines into an asyncio.Queue.
"""
def __init__(
self,
target: str,
queue: asyncio.Queue[str],
loop: asyncio.AbstractEventLoop,
) -> None:
super().__init__()
self._target = target
self._queue = queue
self._loop = loop
self._file = None
self._inode: int | None = None
self._open_target()
def _open_target(self) -> None:
"""
Open the target log file and seek to the end.
"""
try:
self._file = open(self._target, encoding="utf-8", errors="replace")
self._file.seek(0, os.SEEK_END)
self._inode = os.stat(self._target).st_ino
logger.info("Tailing %s (inode %s)", self._target, self._inode)
except FileNotFoundError:
logger.warning("Log file %s not found — waiting for creation", self._target)
self._file = None
self._inode = None
def _read_new_lines(self) -> None:
"""
Read all new complete lines from the current file position.
"""
if self._file is None:
return
for line in self._file:
stripped = line.rstrip("\n\r")
if stripped:
self._loop.call_soon_threadsafe(self._queue.put_nowait, stripped)
def _handle_rotation(self) -> None:
"""
Finish reading the old file, then reopen the target at position 0.
"""
self._read_new_lines()
if self._file is not None:
self._file.close()
try:
self._file = open(self._target, encoding="utf-8", errors="replace")
self._inode = os.stat(self._target).st_ino
logger.info("Rotated to new %s (inode %s)", self._target, self._inode)
except FileNotFoundError:
self._file = None
self._inode = None
def _inode_changed(self) -> bool:
"""
Check whether the target file's inode differs from the one we opened.
"""
try:
current_inode = os.stat(self._target).st_ino
return current_inode != self._inode
except FileNotFoundError:
return False
def on_modified(self, event: FileModifiedEvent) -> None:
"""
Handle new data appended to the log file.
"""
if not isinstance(event, FileModifiedEvent) or event.is_directory:
return
if Path(event.src_path).resolve() != Path(self._target).resolve():
return
if self._inode_changed():
self._handle_rotation()
return
self._read_new_lines()
def on_moved(self, event: FileMovedEvent) -> None:
"""
Handle log rotation via rename (access.log -> access.log.1).
"""
if not isinstance(event, FileMovedEvent):
return
if Path(event.src_path).resolve() == Path(self._target).resolve():
logger.info("Log rotated: %s -> %s", event.src_path, event.dest_path)
self._handle_rotation()
def on_created(self, event: FileCreatedEvent) -> None:
"""
Handle log rotation where a new file is created at the target path.
"""
if not isinstance(event, FileCreatedEvent) or event.is_directory:
return
if Path(event.src_path).resolve() == Path(self._target).resolve():
logger.info("New log file created: %s", event.src_path)
self._handle_rotation()
def close(self) -> None:
"""
Close the underlying file handle.
"""
if self._file is not None:
self._file.close()
self._file = None
class LogTailer:
"""
Watchdog-based nginx log tailer that pushes raw lines
into an asyncio.Queue for downstream processing.
"""
def __init__(
self,
log_path: str,
queue: asyncio.Queue[str],
loop: asyncio.AbstractEventLoop,
) -> None:
self._log_path = log_path
self._handler = _LogHandler(log_path, queue, loop)
self._observer = Observer()
self._started = False
def start(self) -> None:
"""
Begin watching the log file's parent directory for changes.
"""
watch_dir = str(Path(self._log_path).resolve().parent)
self._observer.schedule(self._handler, watch_dir, recursive=False)
self._observer.start()
self._started = True
logger.info("LogTailer started — watching %s", watch_dir)
def stop(self) -> None:
"""
Stop the watchdog observer and close file handles.
"""
if self._started:
self._observer.stop()
self._observer.join(timeout=5)
self._started = False
self._handler.close()
logger.info("LogTailer stopped")
@property
def is_active(self) -> bool:
"""
Whether the tailer is currently running.
"""
return self._started and self._observer.is_alive()

55
PROJECTS/advanced/ai-threat-detection/backend/app/core/redis_manager.py Normal file
View File

@ -0,0 +1,55 @@
"""
©AngelaMos | 2026
redis_manager.py
"""
import redis.asyncio as aioredis
from app.config import settings
class RedisManager:
"""
Async Redis connection lifecycle manager.
"""
def __init__(self) -> None:
self._client: aioredis.Redis | None = None
async def connect(self) -> None:
"""
Create the async Redis client from configured URL.
"""
self._client = aioredis.from_url(
settings.redis_url,
decode_responses=True,
)
async def disconnect(self) -> None:
"""
Close the Redis connection and release resources.
"""
if self._client:
await self._client.aclose()
self._client = None
@property
def client(self) -> aioredis.Redis | None:
"""
Return the active Redis client or None if not connected.
"""
return self._client
async def ping(self) -> bool:
"""
Check Redis connectivity via PING command.
"""
if not self._client:
return False
try:
return await self._client.ping()
except Exception:
return False
redis_manager = RedisManager()

49
PROJECTS/advanced/ai-threat-detection/backend/app/factory.py Normal file
View File

@ -0,0 +1,49 @@
"""
©AngelaMos | 2026
factory.py
"""
import time
from collections.abc import AsyncIterator
from contextlib import asynccontextmanager
from fastapi import FastAPI
from app.config import settings
from app.core.redis_manager import redis_manager
@asynccontextmanager
async def lifespan(app: FastAPI) -> AsyncIterator[None]:
"""
Manage application startup and shutdown lifecycle
"""
app.state.startup_time = time.monotonic()
app.state.pipeline_running = False
await redis_manager.connect()
yield
await redis_manager.disconnect()
def create_app() -> FastAPI:
"""
Build and configure the AngelusVigil FastAPI application
"""
app = FastAPI(
title=settings.app_name,
version="0.1.0",
lifespan=lifespan,
)
app.state.startup_time = time.monotonic()
app.state.pipeline_running = False
from app.api.health import router as health_router
app.include_router(health_router)
return app

8
PROJECTS/advanced/ai-threat-detection/backend/app/main.py Normal file
View File

@ -0,0 +1,8 @@
"""
©AngelaMos | 2026
main.py
"""
from app.factory import create_app
app = create_app()

9
PROJECTS/advanced/ai-threat-detection/backend/app/models/__init__.py Normal file
View File

@ -0,0 +1,9 @@
"""
©AngelaMos | 2026
__init__.py
"""
from app.models.model_metadata import ModelMetadata
from app.models.threat_event import ThreatEvent
__all__ = ["ModelMetadata", "ThreatEvent"]

23
PROJECTS/advanced/ai-threat-detection/backend/app/models/base.py Normal file
View File

@ -0,0 +1,23 @@
"""
©AngelaMos | 2026
base.py
"""
import uuid
from datetime import datetime
from sqlalchemy import DateTime, text
from sqlmodel import Field, SQLModel
class TimestampedModel(SQLModel):
"""
Abstract base providing UUID primary key and timezone-aware created_at.
"""
id: uuid.UUID = Field(default_factory=uuid.uuid4, primary_key=True)
created_at: datetime = Field(
sa_type=DateTime(timezone=True),
sa_column_kwargs={"server_default": text("CURRENT_TIMESTAMP")},
nullable=False,
)

35
PROJECTS/advanced/ai-threat-detection/backend/app/models/model_metadata.py Normal file
View File

@ -0,0 +1,35 @@
"""
©AngelaMos | 2026
model_metadata.py
"""
from sqlalchemy import Column, Index, JSON, text
from sqlmodel import Field
from app.models.base import TimestampedModel
class ModelMetadata(TimestampedModel, table=True):
"""
Tracks ML model versions, training metrics, and deployment status.
"""
__tablename__ = "model_metadata"
__table_args__ = (
Index(
"idx_model_metadata_active",
"model_type",
unique=True,
postgresql_where=text("is_active = TRUE"),
),
)
model_type: str = Field(max_length=30)
version: str = Field(max_length=64)
training_samples: int
metrics: dict = Field(sa_column=Column(JSON, nullable=False))
artifact_path: str
is_active: bool = Field(default=False)
mlflow_run_id: str | None = Field(default=None, max_length=64)
threshold: float | None = Field(default=None)
notes: str | None = Field(default=None)

57
PROJECTS/advanced/ai-threat-detection/backend/app/models/threat_event.py Normal file
View File

@ -0,0 +1,57 @@
"""
©AngelaMos | 2026
threat_event.py
"""
from sqlalchemy import Column, Float, Index, JSON, SmallInteger, text
from sqlmodel import Field
from app.models.base import TimestampedModel
class ThreatEvent(TimestampedModel, table=True):
"""
Primary table for detected threat events.
"""
__tablename__ = "threat_events"
__table_args__ = (
Index("idx_threat_events_created_at", "created_at"),
Index("idx_threat_events_source_ip", "source_ip"),
Index("idx_threat_events_severity", "severity"),
Index("idx_threat_events_score", "threat_score"),
Index(
"idx_threat_events_reviewed",
"reviewed",
postgresql_where=text("reviewed = FALSE"),
),
)
source_ip: str = Field(max_length=45)
request_method: str = Field(max_length=10)
request_path: str
status_code: int = Field(
sa_column=Column(SmallInteger, nullable=False)
)
response_size: int
user_agent: str
threat_score: float = Field(
sa_column=Column(Float, nullable=False)
)
severity: str = Field(max_length=6)
component_scores: dict[str, float] = Field(
sa_column=Column(JSON, nullable=False)
)
geo_country: str | None = Field(default=None, max_length=2)
geo_city: str | None = Field(default=None, max_length=255)
geo_lat: float | None = Field(default=None)
geo_lon: float | None = Field(default=None)
feature_vector: list[float] = Field(
sa_column=Column(JSON, nullable=False)
)
matched_rules: list[str] | None = Field(
default=None, sa_column=Column(JSON, nullable=True)
)
model_version: str | None = Field(default=None, max_length=64)
reviewed: bool = Field(default=False)
review_label: str | None = Field(default=None, max_length=20)

4
PROJECTS/advanced/ai-threat-detection/backend/cli/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

4
PROJECTS/advanced/ai-threat-detection/backend/ml/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

221
PROJECTS/advanced/ai-threat-detection/backend/pyproject.toml Normal file
View File

@ -0,0 +1,221 @@
# ©AngelaMos | 2026
# pyproject.toml
[project]
name = "angelusvigil"
version = "0.1.0"
description = "AI-powered threat detection engine for web server access logs"
requires-python = ">=3.14"
authors = [
{name = "Carter", email = "carter@certgames.com"}
]
dependencies = [
"fastapi>=0.128.8",
"uvicorn[standard]>=0.40.0",
"redis[hiredis]>=7.1.1",
"sqlalchemy>=2.0.46",
"sqlmodel>=0.0.33",
"alembic>=1.18.4",
"asyncpg>=0.31.0",
"pydantic>=2.12.5",
"pydantic-settings>=2.12.0",
"watchdog>=6.0.0",
"geoip2>=5.2.0",
"typer>=0.23.0",
"orjson>=3.11.7",
"httpx>=0.28.1",
]
[project.optional-dependencies]
dev = [
"pytest>=9.0.2",
"pytest-asyncio>=1.3.0",
"aiosqlite>=0.22.1",
"ruff>=0.15.0",
"mypy>=1.19.1",
"pylint>=4.0.4",
"pylint-pydantic>=0.4.1",
"types-redis>=4.6.0.20241004",
"coverage>=7.13.4",
"fakeredis>=2.26.0",
]
ml = [
"torch>=2.10.0",
"scikit-learn>=1.8.0",
"onnxruntime>=1.24.1",
"skl2onnx>=1.20.0",
"mlflow>=3.9.0",
"numpy>=2.4.2",
"pandas>=2.2.0,<3",
"imbalanced-learn>=0.14.1",
]
[project.scripts]
vigil = "cli.main:app"
[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"
[tool.hatch.build.targets.wheel]
packages = ["app", "cli", "ml"]
[tool.ruff]
target-version = "py314"
line-length = 95
indent-width = 4
exclude = [
".bzr",
".direnv",
".eggs",
".git",
".git-rewrite",
".hg",
".ipynb_checkpoints",
".mypy_cache",
".nox",
".pyenv",
".pytest_cache",
".pytype",
".ruff_cache",
".svn",
".tox",
".venv",
"__pycache__",
"_build",
"build",
"dist",
"site-packages",
"venv",
"alembic",
]
[tool.ruff.format]
line-ending = "auto"
skip-magic-trailing-comma = false
[tool.ruff.lint]
select = [
"E1",
"E4",
"E7",
"F",
"W292",
"W605",
"B",
"C4",
"UP",
"ARG",
"SIM",
"I",
"F401",
"F811",
"F821",
]
ignore = [
"E501",
"W291",
"W293",
"I001",
"RUF001",
"RUF002",
"B008",
"ARG001",
"E712",
]
[tool.ruff.lint.per-file-ignores]
"__init__.py" = ["F401"]
"tests/*" = ["ARG002"]
[tool.mypy]
python_version = "3.14"
strict = true
warn_return_any = true
warn_unused_configs = true
disallow_untyped_defs = true
disallow_any_generics = true
check_untyped_defs = true
no_implicit_optional = true
warn_redundant_casts = true
warn_unused_ignores = true
warn_no_return = true
follow_imports = "normal"
ignore_missing_imports = true
[[tool.mypy.overrides]]
module = "app.models.*"
disable_error_code = ["call-arg", "misc"]
[tool.pylint.main]
py-version = "3.12"
jobs = 4
load-plugins = [
"pylint_pydantic",
]
persistent = true
suggestion-mode = true
ignore = [
"venv",
".venv",
"__pycache__",
"build",
"dist",
".git",
".pytest_cache",
".mypy_cache",
".ruff_cache",
"alembic",
]
ignore-paths = [
"^venv/.*",
"^.venv/.*",
"^build/.*",
"^dist/.*",
]
[tool.pylint.messages_control]
disable = [
"C0111",
"C0103",
"R0903",
"W0511",
"W0622",
"W0612",
"W0613",
"C0301",
"C0302",
"C0411",
"C0305",
"C0303",
"C0304",
"R0801",
"E0401",
"C0412",
"W0718",
"E0611",
"E1101",
]
[tool.pylint.design]
max-args = 7
max-attributes = 10
max-locals = 30
max-positional-arguments = 7
[tool.pylint."messages control"]
per-file-ignores = [
"alembic/env.py:W0611",
"app/services/*:C0121",
"tests/*:W0212,W0621",
]
[tool.pytest.ini_options]
asyncio_mode = "auto"
testpaths = ["tests"]
python_files = ["test_*.py"]
python_classes = ["Test*"]
python_functions = ["test_*"]

4
PROJECTS/advanced/ai-threat-detection/backend/tests/__init__.py Normal file
View File

@ -0,0 +1,4 @@
"""
©AngelaMos | 2026
__init__.py
"""

23
PROJECTS/advanced/ai-threat-detection/backend/tests/conftest.py Normal file
View File

@ -0,0 +1,23 @@
"""
©AngelaMos | 2026
conftest.py
"""
import pytest
from app.config import Settings
@pytest.fixture
def test_settings() -> Settings:
"""
Override settings for test environment.
"""
return Settings(
env="testing",
debug=False,
database_url="sqlite+aiosqlite:///test.db",
redis_url="redis://localhost:6379/1",
nginx_log_path="/tmp/test-access.log",
geoip_db_path="/tmp/nonexistent.mmdb",
)

57
PROJECTS/advanced/ai-threat-detection/backend/tests/test_api.py Normal file
View File

@ -0,0 +1,57 @@
"""
©AngelaMos | 2026
test_api.py
"""
import pytest
from httpx import ASGITransport, AsyncClient
from app.main import app
@pytest.mark.asyncio
async def test_health_returns_200() -> None:
"""
Health endpoint returns 200 with status and uptime.
"""
transport = ASGITransport(app=app)
async with AsyncClient(transport=transport, base_url="http://test") as client:
response = await client.get("/health")
assert response.status_code == 200
data = response.json()
assert data["status"] == "healthy"
assert "uptime_seconds" in data
assert isinstance(data["uptime_seconds"], (int, float))
assert "pipeline_running" in data
@pytest.mark.asyncio
async def test_health_returns_pipeline_status() -> None:
"""
Health response includes pipeline_running boolean.
"""
transport = ASGITransport(app=app)
async with AsyncClient(transport=transport, base_url="http://test") as client:
response = await client.get("/health")
data = response.json()
assert isinstance(data["pipeline_running"], bool)
@pytest.mark.asyncio
async def test_ready_returns_check_structure() -> None:
"""
Readiness endpoint returns structured component checks.
"""
transport = ASGITransport(app=app)
async with AsyncClient(transport=transport, base_url="http://test") as client:
response = await client.get("/ready")
assert response.status_code in (200, 503)
data = response.json()
assert "status" in data
assert "checks" in data
assert "database" in data["checks"]
assert "redis" in data["checks"]
assert "models_loaded" in data["checks"]

234
PROJECTS/advanced/ai-threat-detection/backend/tests/test_detection.py Normal file
View File

@ -0,0 +1,234 @@
"""
©AngelaMos | 2026
test_detection.py
"""
from datetime import datetime, timezone
from app.core.detection.rules import RuleEngine, RuleResult
from app.core.ingestion.parsers import ParsedLogEntry
def _make_entry(
path: str = "/api/v1/users",
query_string: str = "",
method: str = "GET",
status_code: int = 200,
user_agent: str = "Mozilla/5.0 (Windows NT 10.0; Win64; x64)",
) -> ParsedLogEntry:
"""
Build a ParsedLogEntry with sensible defaults for rule engine testing.
"""
return ParsedLogEntry(
ip="93.184.216.34",
timestamp=datetime(2026, 2, 11, 14, 30, 0, tzinfo=timezone.utc),
method=method,
path=path,
query_string=query_string,
status_code=status_code,
response_size=1234,
referer="",
user_agent=user_agent,
raw_line="",
)
def _empty_windowed() -> dict[str, float]:
"""
Windowed features representing a quiet IP.
"""
return {
"req_count_1m": 1.0,
"req_count_5m": 3.0,
"req_count_10m": 5.0,
"error_rate_5m": 0.0,
"unique_paths_5m": 2.0,
"unique_uas_10m": 1.0,
"method_entropy_5m": 0.0,
"avg_response_size_5m": 1024.0,
"status_diversity_5m": 1.0,
"path_depth_variance_5m": 0.0,
"inter_request_time_mean": 5000.0,
"inter_request_time_std": 1000.0,
}
ENGINE = RuleEngine()
def test_normal_request_low_severity() -> None:
"""
A benign GET request scores below 0.5 with LOW severity.
"""
result = ENGINE.score_request(_empty_windowed(), _make_entry())
assert result.threat_score < 0.5
assert result.severity == "LOW"
assert result.matched_rules == []
def test_sqli_in_query_string() -> None:
"""
SQL injection payload triggers HIGH severity with SQL_INJECTION rule.
"""
entry = _make_entry(
path="/users",
query_string="id=1' UNION SELECT username,password FROM users--",
)
result = ENGINE.score_request(_empty_windowed(), entry)
assert result.threat_score >= 0.7
assert result.severity == "HIGH"
assert "SQL_INJECTION" in result.matched_rules
def test_xss_in_query_string() -> None:
"""
XSS payload triggers HIGH severity with XSS rule.
"""
entry = _make_entry(
path="/comment",
query_string="body=<script>alert(document.cookie)</script>",
)
result = ENGINE.score_request(_empty_windowed(), entry)
assert result.threat_score >= 0.7
assert result.severity == "HIGH"
assert "XSS" in result.matched_rules
def test_path_traversal() -> None:
"""
Path traversal triggers at least MEDIUM severity.
"""
entry = _make_entry(path="/static/../../etc/passwd")
result = ENGINE.score_request(_empty_windowed(), entry)
assert result.threat_score >= 0.5
assert result.severity in {"MEDIUM", "HIGH"}
assert "PATH_TRAVERSAL" in result.matched_rules
def test_command_injection() -> None:
"""
Command injection triggers HIGH severity.
"""
entry = _make_entry(
path="/ping",
query_string="host=127.0.0.1;cat /etc/passwd",
)
result = ENGINE.score_request(_empty_windowed(), entry)
assert result.threat_score >= 0.7
assert result.severity == "HIGH"
assert "COMMAND_INJECTION" in result.matched_rules
def test_scanner_ua_contributes() -> None:
"""
Known scanner UA fires the SCANNER_UA rule.
"""
entry = _make_entry(user_agent="Mozilla/5.00 (Nikto/2.1.6)")
result = ENGINE.score_request(_empty_windowed(), entry)
assert "SCANNER_UA" in result.matched_rules
assert result.threat_score > 0.0
def test_rate_anomaly() -> None:
"""
High request rate in the 1-minute window fires RATE_ANOMALY.
"""
windowed = _empty_windowed()
windowed["req_count_1m"] = 150.0
result = ENGINE.score_request(windowed, _make_entry())
assert "RATE_ANOMALY" in result.matched_rules
assert result.threat_score > 0.0
def test_multiple_rules_aggregate() -> None:
"""
Multiple triggered rules produce a higher score than any single rule alone.
"""
entry = _make_entry(
path="/users",
query_string="id=1' OR 1=1--",
user_agent="sqlmap/1.8",
)
sqli_only = ENGINE.score_request(
_empty_windowed(),
_make_entry(path="/users", query_string="id=1' OR 1=1--"),
)
combined = ENGINE.score_request(_empty_windowed(), entry)
assert combined.threat_score > sqli_only.threat_score
assert len(combined.matched_rules) > len(sqli_only.matched_rules)
def test_score_clamped_to_unit_range() -> None:
"""
Score never exceeds 1.0 even with many triggered rules.
"""
entry = _make_entry(
path="/static/../../etc/passwd;cat /etc/shadow",
query_string="id=1' UNION SELECT 1--&x=<script>alert(1)</script>",
user_agent="sqlmap/1.8",
)
windowed = _empty_windowed()
windowed["req_count_1m"] = 200.0
windowed["error_rate_5m"] = 0.8
result = ENGINE.score_request(windowed, entry)
assert 0.0 <= result.threat_score <= 1.0
def test_severity_thresholds() -> None:
"""
Severity classification follows the architecture thresholds.
"""
low = ENGINE.score_request(_empty_windowed(), _make_entry())
assert low.severity == "LOW"
medium = ENGINE.score_request(
_empty_windowed(),
_make_entry(path="/static/../../etc/passwd"),
)
assert medium.severity in {"MEDIUM", "HIGH"}
high = ENGINE.score_request(
_empty_windowed(),
_make_entry(
path="/users",
query_string="id=1' UNION SELECT username FROM users--",
),
)
assert high.severity == "HIGH"
def test_result_has_component_scores() -> None:
"""
RuleResult includes individual scores per triggered rule.
"""
entry = _make_entry(
path="/users",
query_string="id=1' OR 1=1--",
)
result = ENGINE.score_request(_empty_windowed(), entry)
assert len(result.component_scores) == len(result.matched_rules)
for rule_name in result.matched_rules:
assert rule_name in result.component_scores
assert result.component_scores[rule_name] > 0.0
def test_file_inclusion() -> None:
"""
PHP stream wrapper in query string triggers FILE_INCLUSION rule.
"""
entry = _make_entry(
path="/include",
query_string="page=php://filter/convert.base64-encode/resource=config",
)
result = ENGINE.score_request(_empty_windowed(), entry)
assert "FILE_INCLUSION" in result.matched_rules
assert result.threat_score >= 0.5
def test_double_encoding() -> None:
"""
Double-encoded characters trigger the DOUBLE_ENCODING rule.
"""
entry = _make_entry(path="/path%2527trick")
result = ENGINE.score_request(_empty_windowed(), entry)
assert "DOUBLE_ENCODING" in result.matched_rules

592
PROJECTS/advanced/ai-threat-detection/backend/tests/test_features.py Normal file
View File

@ -0,0 +1,592 @@
"""
©AngelaMos | 2026
test_features.py
"""
from datetime import datetime, timezone
from app.core.features.extractor import extract_request_features
from app.core.ingestion.parsers import ParsedLogEntry
FEATURE_KEYS = {
"http_method",
"path_depth",
"path_entropy",
"path_length",
"query_string_length",
"query_param_count",
"has_encoded_chars",
"has_double_encoding",
"status_code",
"status_class",
"response_size",
"hour_of_day",
"day_of_week",
"is_weekend",
"ua_length",
"ua_entropy",
"is_known_bot",
"is_known_scanner",
"has_attack_pattern",
"special_char_ratio",
"file_extension",
"country_code",
"is_private_ip",
}
def _make_entry(
ip: str = "93.184.216.34",
timestamp: datetime | None = None,
method: str = "GET",
path: str = "/api/v1/users",
query_string: str = "",
status_code: int = 200,
response_size: int = 1234,
referer: str = "",
user_agent: str = "Mozilla/5.0 (Windows NT 10.0; Win64; x64)",
) -> ParsedLogEntry:
"""
Build a ParsedLogEntry with sensible defaults for testing.
"""
if timestamp is None:
timestamp = datetime(2026, 2, 11, 14, 30, 0, tzinfo=timezone.utc)
return ParsedLogEntry(
ip=ip,
timestamp=timestamp,
method=method,
path=path,
query_string=query_string,
status_code=status_code,
response_size=response_size,
referer=referer,
user_agent=user_agent,
raw_line="",
)
def test_extract_returns_all_23_feature_keys() -> None:
"""
Feature dict contains exactly the 23 per-request keys.
"""
features = extract_request_features(_make_entry())
assert set(features.keys()) == FEATURE_KEYS
def test_path_depth() -> None:
"""
Path depth counts non-empty segments between slashes.
"""
assert extract_request_features(_make_entry(path="/"))["path_depth"] == 0
assert extract_request_features(_make_entry(path="/api"))["path_depth"] == 1
assert extract_request_features(
_make_entry(path="/api/v1/users")
)["path_depth"] == 3
def test_path_entropy_high_vs_low() -> None:
"""
Random-character paths have higher entropy than simple paths.
"""
low = extract_request_features(
_make_entry(path="/index.html")
)["path_entropy"]
high = extract_request_features(
_make_entry(path="/x8Kp2mQz7wR4vL1n")
)["path_entropy"]
assert high > low
def test_query_string_features() -> None:
"""
Query param count and length are extracted correctly.
"""
features = extract_request_features(
_make_entry(query_string="page=1&sort=name&limit=50")
)
assert features["query_param_count"] == 3
assert features["query_string_length"] == len("page=1&sort=name&limit=50")
empty = extract_request_features(_make_entry(query_string=""))
assert empty["query_param_count"] == 0
assert empty["query_string_length"] == 0
def test_url_encoding_detection() -> None:
"""
Percent-encoded sequences are detected in path and query.
"""
encoded = extract_request_features(
_make_entry(path="/search", query_string="q=%27OR+1%3D1")
)
assert encoded["has_encoded_chars"] is True
clean = extract_request_features(
_make_entry(path="/index.html", query_string="")
)
assert clean["has_encoded_chars"] is False
def test_double_encoding_detection() -> None:
"""
Double-encoded sequences like %2527 are flagged.
"""
double = extract_request_features(
_make_entry(path="/path%2527trick")
)
assert double["has_double_encoding"] is True
single = extract_request_features(
_make_entry(path="/path%27normal")
)
assert single["has_double_encoding"] is False
def test_status_class() -> None:
"""
Status class groups status codes into Nxx buckets.
"""
assert extract_request_features(
_make_entry(status_code=200)
)["status_class"] == "2xx"
assert extract_request_features(
_make_entry(status_code=404)
)["status_class"] == "4xx"
assert extract_request_features(
_make_entry(status_code=503)
)["status_class"] == "5xx"
def test_temporal_features() -> None:
"""
Hour, day of week, and weekend flag derived from timestamp.
"""
wednesday_2pm = datetime(2026, 2, 11, 14, 0, 0, tzinfo=timezone.utc)
features = extract_request_features(
_make_entry(timestamp=wednesday_2pm)
)
assert features["hour_of_day"] == 14
assert features["day_of_week"] == 2
assert features["is_weekend"] is False
saturday_3am = datetime(2026, 2, 14, 3, 0, 0, tzinfo=timezone.utc)
weekend = extract_request_features(
_make_entry(timestamp=saturday_3am)
)
assert weekend["hour_of_day"] == 3
assert weekend["day_of_week"] == 5
assert weekend["is_weekend"] is True
def test_ua_bot_detection() -> None:
"""
Known bot user agents are flagged.
"""
bot = extract_request_features(
_make_entry(
user_agent="Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)"
)
)
assert bot["is_known_bot"] is True
normal = extract_request_features(
_make_entry(user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64)")
)
assert normal["is_known_bot"] is False
def test_ua_scanner_detection() -> None:
"""
Known vulnerability scanner user agents are flagged.
"""
nikto = extract_request_features(
_make_entry(user_agent="Mozilla/5.00 (Nikto/2.1.6)")
)
assert nikto["is_known_scanner"] is True
sqlmap = extract_request_features(
_make_entry(user_agent="sqlmap/1.8")
)
assert sqlmap["is_known_scanner"] is True
def test_attack_pattern_detection() -> None:
"""
SQLi, XSS, and path traversal patterns in paths are detected.
"""
sqli = extract_request_features(
_make_entry(path="/users", query_string="id=1' OR 1=1--")
)
assert sqli["has_attack_pattern"] is True
xss = extract_request_features(
_make_entry(path="/comment", query_string="body=<script>alert(1)</script>")
)
assert xss["has_attack_pattern"] is True
traversal = extract_request_features(
_make_entry(path="/static/../../etc/passwd")
)
assert traversal["has_attack_pattern"] is True
clean = extract_request_features(
_make_entry(path="/api/v1/health")
)
assert clean["has_attack_pattern"] is False
def test_special_char_ratio() -> None:
"""
Paths with many non-alphanumeric characters have higher ratios.
"""
clean = extract_request_features(
_make_entry(path="/api/users")
)["special_char_ratio"]
noisy = extract_request_features(
_make_entry(path="/<script>alert('xss')</script>")
)["special_char_ratio"]
assert noisy > clean
def test_private_ip_detection() -> None:
"""
RFC 1918 and loopback addresses are classified as private.
"""
assert extract_request_features(
_make_entry(ip="192.168.1.1")
)["is_private_ip"] is True
assert extract_request_features(
_make_entry(ip="127.0.0.1")
)["is_private_ip"] is True
assert extract_request_features(
_make_entry(ip="8.8.8.8")
)["is_private_ip"] is False
def test_file_extension() -> None:
"""
File extension is extracted from the path.
"""
assert extract_request_features(
_make_entry(path="/style.css")
)["file_extension"] == ".css"
assert extract_request_features(
_make_entry(path="/api/users")
)["file_extension"] == ""
def test_country_code_passthrough() -> None:
"""
Country code is passed through from the caller.
"""
features = extract_request_features(
_make_entry(), country_code="US"
)
assert features["country_code"] == "US"
features_empty = extract_request_features(_make_entry())
assert features_empty["country_code"] == ""
import time
import fakeredis.aioredis
import pytest
from app.core.features.aggregator import WindowAggregator
AGGREGATOR_KEYS = {
"req_count_1m",
"req_count_5m",
"req_count_10m",
"error_rate_5m",
"unique_paths_5m",
"unique_uas_10m",
"method_entropy_5m",
"avg_response_size_5m",
"status_diversity_5m",
"path_depth_variance_5m",
"inter_request_time_mean",
"inter_request_time_std",
}
@pytest.fixture
async def aggregator():
"""
WindowAggregator backed by an in-memory fake Redis.
"""
redis = fakeredis.aioredis.FakeRedis(decode_responses=True)
yield WindowAggregator(redis)
await redis.aclose()
async def _record(
agg: WindowAggregator,
ip: str = "10.0.0.1",
request_id: str = "r-001",
path: str = "/api/users",
path_depth: int = 2,
method: str = "GET",
status_code: int = 200,
user_agent: str = "Mozilla/5.0",
response_size: int = 1024,
timestamp: float | None = None,
) -> dict[str, float]:
"""
Shorthand for record_and_aggregate with sensible defaults.
"""
return await agg.record_and_aggregate(
ip=ip,
request_id=request_id,
path=path,
path_depth=path_depth,
method=method,
status_code=status_code,
user_agent=user_agent,
response_size=response_size,
timestamp=timestamp or time.time(),
)
@pytest.mark.asyncio
async def test_aggregator_single_request(aggregator) -> None:
"""
A single request yields count of 1.
"""
result = await _record(aggregator)
assert set(result.keys()) == AGGREGATOR_KEYS
assert result["req_count_1m"] == 1
assert result["req_count_5m"] == 1
assert result["req_count_10m"] == 1
@pytest.mark.asyncio
async def test_aggregator_ten_requests(aggregator) -> None:
"""
Ten requests within 30 seconds all count in the 1-minute window.
"""
now = time.time()
result = None
for i in range(10):
result = await _record(
aggregator,
request_id=f"r-{i:03d}",
timestamp=now - 30 + i * 3,
)
assert result is not None
assert result["req_count_1m"] == 10
@pytest.mark.asyncio
async def test_aggregator_error_rate(aggregator) -> None:
"""
Error rate is the ratio of 4xx/5xx responses.
"""
now = time.time()
for i in range(8):
await _record(
aggregator,
request_id=f"ok-{i}",
status_code=200,
timestamp=now - 60 + i,
)
for i in range(2):
result = await _record(
aggregator,
request_id=f"err-{i}",
status_code=404,
timestamp=now - 10 + i,
)
assert result["error_rate_5m"] == pytest.approx(0.2, abs=0.01)
@pytest.mark.asyncio
async def test_aggregator_unique_paths(aggregator) -> None:
"""
Unique paths counts distinct URL paths in the window.
"""
now = time.time()
paths = ["/api/users", "/api/posts", "/api/users", "/api/health"]
result = None
for i, p in enumerate(paths):
result = await _record(
aggregator,
request_id=f"r-{i}",
path=p,
timestamp=now - 10 + i,
)
assert result is not None
assert result["unique_paths_5m"] == 3
@pytest.mark.asyncio
async def test_aggregator_unique_uas(aggregator) -> None:
"""
Unique UAs counts distinct user agents in the 10-minute window.
"""
now = time.time()
uas = ["Mozilla/5.0", "curl/8.0", "Mozilla/5.0", "python-httpx/0.28"]
result = None
for i, ua in enumerate(uas):
result = await _record(
aggregator,
request_id=f"r-{i}",
user_agent=ua,
timestamp=now - 10 + i,
)
assert result is not None
assert result["unique_uas_10m"] == 3
@pytest.mark.asyncio
async def test_aggregator_ttl_set(aggregator) -> None:
"""
All Redis keys are set with a 900-second TTL.
"""
await _record(aggregator, ip="5.5.5.5")
ttl = await aggregator._redis.ttl("ip:5.5.5.5:requests")
assert 0 < ttl <= 900
@pytest.mark.asyncio
async def test_aggregator_window_boundary(aggregator) -> None:
"""
Requests outside the 1-minute window are excluded from req_count_1m
but still counted in req_count_5m.
"""
now = time.time()
await _record(
aggregator,
request_id="old",
timestamp=now - 120,
)
result = await _record(
aggregator,
request_id="new",
timestamp=now,
)
assert result["req_count_1m"] == 1
assert result["req_count_5m"] == 2
from app.core.features.encoder import encode_for_inference
from app.core.features.mappings import FEATURE_ORDER, METHOD_MAP, STATUS_CLASS_MAP
def _full_features() -> dict[str, int | float | bool | str]:
"""
Build a complete 35-key feature dict with realistic values.
"""
return {
"http_method": "GET",
"path_depth": 3,
"path_entropy": 3.12,
"path_length": 14,
"query_string_length": 6,
"query_param_count": 1,
"has_encoded_chars": False,
"has_double_encoding": False,
"status_code": 200,
"status_class": "2xx",
"response_size": 1234,
"hour_of_day": 14,
"day_of_week": 2,
"is_weekend": False,
"ua_length": 42,
"ua_entropy": 4.01,
"is_known_bot": False,
"is_known_scanner": False,
"has_attack_pattern": False,
"special_char_ratio": 0.21,
"file_extension": ".html",
"country_code": "US",
"is_private_ip": False,
"req_count_1m": 5.0,
"req_count_5m": 20.0,
"req_count_10m": 45.0,
"error_rate_5m": 0.1,
"unique_paths_5m": 8.0,
"unique_uas_10m": 2.0,
"method_entropy_5m": 0.5,
"avg_response_size_5m": 2048.0,
"status_diversity_5m": 3.0,
"path_depth_variance_5m": 1.2,
"inter_request_time_mean": 250.0,
"inter_request_time_std": 80.0,
}
def test_encoder_output_shape_and_type() -> None:
"""
Encoded vector has exactly 35 float elements.
"""
result = encode_for_inference(_full_features())
assert len(result) == 35
assert all(isinstance(v, float) for v in result)
def test_encoder_method_ordinal() -> None:
"""
HTTP methods map to deterministic ordinal indices.
"""
features = _full_features()
features["http_method"] = "GET"
vec = encode_for_inference(features)
assert vec[FEATURE_ORDER.index("http_method")] == float(METHOD_MAP["GET"])
features["http_method"] = "POST"
vec = encode_for_inference(features)
assert vec[FEATURE_ORDER.index("http_method")] == float(METHOD_MAP["POST"])
def test_encoder_status_class_ordinal() -> None:
"""
Status classes map to deterministic ordinal indices.
"""
features = _full_features()
features["status_class"] = "4xx"
vec = encode_for_inference(features)
idx = FEATURE_ORDER.index("status_class")
assert vec[idx] == float(STATUS_CLASS_MAP["4xx"])
def test_encoder_boolean_to_float() -> None:
"""
Boolean features encode to 0.0 or 1.0.
"""
features = _full_features()
features["is_known_bot"] = True
features["is_weekend"] = False
vec = encode_for_inference(features)
assert vec[FEATURE_ORDER.index("is_known_bot")] == 1.0
assert vec[FEATURE_ORDER.index("is_weekend")] == 0.0
def test_encoder_numerical_passthrough() -> None:
"""
Numerical features pass through as raw float values.
"""
features = _full_features()
vec = encode_for_inference(features)
assert vec[FEATURE_ORDER.index("path_depth")] == 3.0
assert vec[FEATURE_ORDER.index("response_size")] == 1234.0
assert vec[FEATURE_ORDER.index("req_count_1m")] == 5.0
def test_encoder_unknown_categorical() -> None:
"""
Unknown categorical values fall back to 0.
"""
features = _full_features()
features["http_method"] = "BREW"
features["status_class"] = "9xx"
vec = encode_for_inference(features)
assert vec[FEATURE_ORDER.index("http_method")] == 0.0
assert vec[FEATURE_ORDER.index("status_class")] == 0.0

143
PROJECTS/advanced/ai-threat-detection/backend/tests/test_geoip.py Normal file
View File

@ -0,0 +1,143 @@
"""
©AngelaMos | 2026
test_geoip.py
"""
import asyncio
from unittest.mock import MagicMock, patch
import pytest
from app.core.enrichment.geoip import GeoIPService, GeoResult
def _mock_city_response(
country_iso: str = "US",
city_name: str = "Los Angeles",
lat: float = 34.0522,
lon: float = -118.2437,
) -> MagicMock:
"""
Build a mock geoip2 City response object.
"""
response = MagicMock()
response.country.iso_code = country_iso
response.city.name = city_name
response.location.latitude = lat
response.location.longitude = lon
return response
@pytest.fixture
def mock_reader() -> MagicMock:
"""
A mock geoip2 database reader.
"""
reader = MagicMock()
reader.city.return_value = _mock_city_response()
return reader
def test_geo_result_fields() -> None:
"""
GeoResult stores country, city, latitude, and longitude.
"""
result = GeoResult(country="US", city="Seattle", lat=47.6, lon=-122.3)
assert result.country == "US"
assert result.city == "Seattle"
assert result.lat == 47.6
assert result.lon == -122.3
@pytest.mark.asyncio
async def test_lookup_returns_geo_result(mock_reader) -> None:
"""
Successful lookup returns a populated GeoResult.
"""
service = GeoIPService.__new__(GeoIPService)
service._reader = mock_reader
result = await service.lookup("8.8.8.8")
assert result is not None
assert result.country == "US"
assert result.city == "Los Angeles"
assert result.lat == pytest.approx(34.0522)
assert result.lon == pytest.approx(-118.2437)
@pytest.mark.asyncio
async def test_lookup_private_ip_returns_none(mock_reader) -> None:
"""
Private/loopback IPs return None without hitting the reader.
"""
service = GeoIPService.__new__(GeoIPService)
service._reader = mock_reader
assert await service.lookup("192.168.1.1") is None
assert await service.lookup("127.0.0.1") is None
assert await service.lookup("10.0.0.1") is None
mock_reader.city.assert_not_called()
@pytest.mark.asyncio
async def test_lookup_address_not_found_returns_none(mock_reader) -> None:
"""
AddressNotFoundError from the reader returns None.
"""
from geoip2.errors import AddressNotFoundError
mock_reader.city.side_effect = AddressNotFoundError("8.8.8.8")
service = GeoIPService.__new__(GeoIPService)
service._reader = mock_reader
assert await service.lookup("8.8.8.8") is None
@pytest.mark.asyncio
async def test_lookup_no_reader_returns_none() -> None:
"""
When no .mmdb file is available, all lookups return None.
"""
service = GeoIPService.__new__(GeoIPService)
service._reader = None
assert await service.lookup("8.8.8.8") is None
@pytest.mark.asyncio
async def test_lookup_missing_city_name(mock_reader) -> None:
"""
Responses with None city name are handled gracefully.
"""
mock_reader.city.return_value = _mock_city_response(
city_name=None, lat=0.0, lon=0.0
)
service = GeoIPService.__new__(GeoIPService)
service._reader = mock_reader
result = await service.lookup("1.1.1.1")
assert result is not None
assert result.country == "US"
assert result.city is None
def test_service_init_missing_db(tmp_path) -> None:
"""
Initializing with a non-existent .mmdb file sets reader to None.
"""
service = GeoIPService(db_path=str(tmp_path / "missing.mmdb"))
assert service._reader is None
def test_service_init_valid_db(tmp_path) -> None:
"""
Initializing with a valid .mmdb path opens the reader.
"""
fake_mmdb = tmp_path / "GeoLite2-City.mmdb"
fake_mmdb.touch()
with patch("app.core.enrichment.geoip.geoip2.database.Reader") as mock_cls:
mock_cls.return_value = MagicMock()
service = GeoIPService(db_path=str(fake_mmdb))
assert service._reader is not None
mock_cls.assert_called_once_with(str(fake_mmdb))

128
PROJECTS/advanced/ai-threat-detection/backend/tests/test_parsers.py Normal file
View File

@ -0,0 +1,128 @@
"""
©AngelaMos | 2026
test_parsers.py
"""
from datetime import datetime, timezone
from app.core.ingestion.parsers import ParsedLogEntry, parse_combined
def test_parse_standard_combined_line() -> None:
"""
Parse a standard nginx combined log line into all fields.
"""
line = (
'93.184.216.34 - - [11/Feb/2026:14:30:00 +0000] '
'"GET /api/users?page=1 HTTP/1.1" 200 1234 '
'"https://example.com" '
'"Mozilla/5.0 (Windows NT 10.0; Win64; x64)"'
)
result = parse_combined(line)
assert result is not None
assert isinstance(result, ParsedLogEntry)
assert result.ip == "93.184.216.34"
assert result.timestamp == datetime(2026, 2, 11, 14, 30, 0, tzinfo=timezone.utc)
assert result.method == "GET"
assert result.path == "/api/users"
assert result.query_string == "page=1"
assert result.status_code == 200
assert result.response_size == 1234
assert result.referer == "https://example.com"
assert result.user_agent == "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"
assert result.raw_line == line
def test_parse_ipv6_address() -> None:
"""
Parse a line with an IPv6 source address.
"""
line = (
'::1 - - [11/Feb/2026:14:30:00 +0000] '
'"GET / HTTP/1.1" 200 612 "-" "curl/8.0"'
)
result = parse_combined(line)
assert result is not None
assert result.ip == "::1"
assert result.method == "GET"
assert result.path == "/"
assert result.query_string == ""
def test_parse_missing_referer() -> None:
"""
A dash referer is normalized to an empty string.
"""
line = (
'10.0.0.1 - - [11/Feb/2026:08:15:42 +0000] '
'"POST /login HTTP/1.1" 302 0 "-" "Mozilla/5.0"'
)
result = parse_combined(line)
assert result is not None
assert result.referer == ""
assert result.method == "POST"
assert result.status_code == 302
def test_parse_complex_query_string() -> None:
"""
Query strings with multiple parameters and special characters.
"""
line = (
'93.184.216.34 - - [11/Feb/2026:14:30:00 +0000] '
'"GET /search?q=hello+world&lang=en&page=2&sort=relevance HTTP/1.1" '
'200 5678 "https://example.com/search" "Mozilla/5.0"'
)
result = parse_combined(line)
assert result is not None
assert result.path == "/search"
assert result.query_string == "q=hello+world&lang=en&page=2&sort=relevance"
def test_parse_malformed_line_returns_none() -> None:
"""
Malformed lines return None instead of raising.
"""
assert parse_combined("this is not a valid log line") is None
def test_parse_empty_line_returns_none() -> None:
"""
Empty input returns None.
"""
assert parse_combined("") is None
def test_parse_dash_response_size() -> None:
"""
A dash response size (e.g. HEAD 304) is normalized to zero.
"""
line = (
'1.2.3.4 - - [11/Feb/2026:10:00:00 +0000] '
'"HEAD / HTTP/1.1" 304 - "-" "Mozilla/5.0"'
)
result = parse_combined(line)
assert result is not None
assert result.response_size == 0
assert result.status_code == 304
def test_parse_full_ipv6_address() -> None:
"""
Parse a line with a full-length IPv6 address.
"""
line = (
'2001:0db8:85a3:0000:0000:8a2e:0370:7334 - - '
'[11/Feb/2026:14:30:00 +0000] "GET /api/v1/health HTTP/2.0" '
'200 256 "-" "python-httpx/0.28"'
)
result = parse_combined(line)
assert result is not None
assert result.ip == "2001:0db8:85a3:0000:0000:8a2e:0370:7334"
assert result.path == "/api/v1/health"

139
PROJECTS/advanced/ai-threat-detection/backend/tests/test_pipeline.py Normal file
View File

@ -0,0 +1,139 @@
"""
©AngelaMos | 2026
test_pipeline.py
"""
import asyncio
import fakeredis.aioredis
import pytest
from app.core.detection.rules import RuleEngine
from app.core.ingestion.pipeline import Pipeline, ScoredRequest
VALID_LINE = (
'93.184.216.34 - - [11/Feb/2026:14:30:00 +0000] '
'"GET /api/v1/users HTTP/1.1" 200 1234 '
'"https://example.com" '
'"Mozilla/5.0 (Windows NT 10.0; Win64; x64)"'
)
SQLI_LINE = (
'93.184.216.34 - - [11/Feb/2026:14:30:01 +0000] '
'"GET /users?id=1%27OR%201=1-- HTTP/1.1" 200 512 '
'"-" "Mozilla/5.0"'
)
async def _make_pipeline(
results: list[ScoredRequest],
raw_queue_size: int = 100,
parsed_queue_size: int = 100,
feature_queue_size: int = 100,
alert_queue_size: int = 100,
) -> Pipeline:
"""
Build a Pipeline wired to an in-memory Redis and a results collector.
"""
redis = fakeredis.aioredis.FakeRedis(decode_responses=True)
async def collect(sr: ScoredRequest) -> None:
results.append(sr)
pipeline = Pipeline(
redis_client=redis,
rule_engine=RuleEngine(),
on_result=collect,
raw_queue_size=raw_queue_size,
parsed_queue_size=parsed_queue_size,
feature_queue_size=feature_queue_size,
alert_queue_size=alert_queue_size,
)
await pipeline.start()
return pipeline
@pytest.mark.asyncio
async def test_end_to_end_valid_line() -> None:
"""
A valid log line flows through all 4 stages and produces a scored result.
"""
results: list[ScoredRequest] = []
pipeline = await _make_pipeline(results)
await pipeline.raw_queue.put(VALID_LINE)
await pipeline.stop()
assert len(results) == 1
assert results[0].entry.ip == "93.184.216.34"
assert results[0].entry.method == "GET"
assert len(results[0].feature_vector) == 35
assert results[0].rule_result.severity == "LOW"
@pytest.mark.asyncio
async def test_malformed_line_dropped() -> None:
"""
Malformed lines are dropped at the parse stage without crashing.
"""
results: list[ScoredRequest] = []
pipeline = await _make_pipeline(results)
await pipeline.raw_queue.put("this is not a valid log line")
await pipeline.raw_queue.put(VALID_LINE)
await pipeline.stop()
assert len(results) == 1
@pytest.mark.asyncio
async def test_backpressure_with_tiny_queues() -> None:
"""
Items flow through correctly even when all queues have maxsize=1.
"""
results: list[ScoredRequest] = []
pipeline = await _make_pipeline(
results,
raw_queue_size=1,
parsed_queue_size=1,
feature_queue_size=1,
alert_queue_size=1,
)
for _ in range(5):
await pipeline.raw_queue.put(VALID_LINE)
await pipeline.stop()
assert len(results) == 5
@pytest.mark.asyncio
async def test_shutdown_drains_and_exits() -> None:
"""
Calling stop() drains remaining items and all tasks exit cleanly.
"""
results: list[ScoredRequest] = []
pipeline = await _make_pipeline(results)
await pipeline.raw_queue.put(VALID_LINE)
await pipeline.raw_queue.put(VALID_LINE)
await pipeline.stop()
assert len(results) == 2
assert all(t.done() for t in pipeline._tasks)
@pytest.mark.asyncio
async def test_attack_line_scored_high() -> None:
"""
A SQLi payload in the log line is scored as HIGH by the rule engine.
"""
results: list[ScoredRequest] = []
pipeline = await _make_pipeline(results)
await pipeline.raw_queue.put(SQLI_LINE)
await pipeline.stop()
assert len(results) == 1
assert results[0].rule_result.severity == "HIGH"
assert "SQL_INJECTION" in results[0].rule_result.matched_rules

2791
PROJECTS/advanced/ai-threat-detection/backend/uv.lock Normal file
View File

File diff suppressed because it is too large Load Diff

110
PROJECTS/advanced/ai-threat-detection/compose.yml Normal file
View File

@ -0,0 +1,110 @@
# ©AngelaMos | 2026
# Production Docker Compose
services:
postgres:
image: postgres:18-alpine
container_name: vigil-postgres
environment:
POSTGRES_DB: ${POSTGRES_DB:-angelusvigil}
POSTGRES_USER: ${POSTGRES_USER:-vigil}
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:?POSTGRES_PASSWORD required}
POSTGRES_INITDB_ARGS: "-E UTF8 --locale=C"
volumes:
- postgres_data:/var/lib/postgresql/data
networks:
- vigil_network
healthcheck:
test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-vigil} -d ${POSTGRES_DB:-angelusvigil}"]
interval: 10s
timeout: 5s
retries: 5
start_period: 30s
restart: always
redis:
image: redis:7.4-alpine
container_name: vigil-redis
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
- redis_data:/data
- ./infra/redis/redis.conf:/usr/local/etc/redis/redis.conf:ro
networks:
- vigil_network
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 10s
timeout: 5s
retries: 5
start_period: 5s
restart: always
backend:
build:
context: .
dockerfile: infra/docker/fastapi.prod
container_name: vigil-backend
environment:
ENV: ${ENV:-production}
DEBUG: ${DEBUG:-false}
LOG_LEVEL: ${LOG_LEVEL:-INFO}
API_KEY: ${API_KEY:?API_KEY required}
DATABASE_URL: postgresql+asyncpg://${POSTGRES_USER:-vigil}:${POSTGRES_PASSWORD}@postgres:5432/${POSTGRES_DB:-angelusvigil}
REDIS_URL: ${REDIS_URL:-redis://redis:6379}
NGINX_LOG_PATH: /var/log/nginx/access.log
GEOIP_DB_PATH: /usr/share/GeoIP/GeoLite2-City.mmdb
volumes:
- nginx_logs:/var/log/nginx:ro
- geoip_data:/usr/share/GeoIP:ro
- model_data:/app/data/models
expose:
- "8000"
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
networks:
- vigil_network
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
start_period: 10s
restart: always
geoip-updater:
image: maxmindinc/geoipupdate:latest
container_name: vigil-geoip
environment:
GEOIPUPDATE_ACCOUNT_ID: ${GEOIP_ACCOUNT_ID:?GEOIP_ACCOUNT_ID required}
GEOIPUPDATE_LICENSE_KEY: ${GEOIP_LICENSE_KEY:?GEOIP_LICENSE_KEY required}
GEOIPUPDATE_EDITION_IDS: "GeoLite2-City"
GEOIPUPDATE_FREQUENCY: "168"
volumes:
- geoip_data:/usr/share/GeoIP
networks:
- vigil_network
restart: always
networks:
vigil_network:
driver: bridge
name: vigil_network
certgames_net:
external: true
name: certgames_default
volumes:
postgres_data:
name: vigil_postgres_data
redis_data:
name: vigil_redis_data
geoip_data:
name: vigil_geoip_data
model_data:
name: vigil_model_data
nginx_logs:
external: true
name: certgames_nginx_logs

0
PROJECTS/advanced/ai-threat-detection/data/sample-logs/.gitkeep Normal file
View File

97
PROJECTS/advanced/ai-threat-detection/dev.compose.yml Normal file
View File

@ -0,0 +1,97 @@
# ©AngelaMos | 2026
# Development Docker Compose
services:
postgres:
image: postgres:18-alpine
container_name: vigil-postgres-dev
environment:
POSTGRES_DB: ${POSTGRES_DB:-angelusvigil}
POSTGRES_USER: ${POSTGRES_USER:-vigil}
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:?POSTGRES_PASSWORD required}
POSTGRES_INITDB_ARGS: "-E UTF8 --locale=C"
volumes:
- postgres_dev_data:/var/lib/postgresql/data
ports:
- "${POSTGRES_HOST_PORT:-16969}:5432"
networks:
- vigil_network_dev
healthcheck:
test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-vigil} -d ${POSTGRES_DB:-angelusvigil}"]
interval: 10s
timeout: 5s
retries: 5
redis:
image: redis:7.4-alpine
container_name: vigil-redis-dev
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
- redis_dev_data:/data
- ./infra/redis/redis.conf:/usr/local/etc/redis/redis.conf:ro
ports:
- "${REDIS_HOST_PORT:-26969}:6379"
networks:
- vigil_network_dev
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 10s
timeout: 5s
retries: 5
backend:
build:
context: .
dockerfile: infra/docker/fastapi.dev
container_name: vigil-backend-dev
environment:
ENV: ${ENV:-development}
DEBUG: ${DEBUG:-true}
LOG_LEVEL: ${LOG_LEVEL:-DEBUG}
API_KEY: ${API_KEY:-dev-api-key}
DATABASE_URL: postgresql+asyncpg://${POSTGRES_USER:-vigil}:${POSTGRES_PASSWORD}@postgres:5432/${POSTGRES_DB:-angelusvigil}
REDIS_URL: ${REDIS_URL:-redis://redis:6379}
NGINX_LOG_PATH: /var/log/nginx/access.log
GEOIP_DB_PATH: /usr/share/GeoIP/GeoLite2-City.mmdb
volumes:
- ./backend:/app
- /app/.venv
- /app/__pycache__
- ./data/sample-logs:/var/log/nginx:ro
- geoip_dev_data:/usr/share/GeoIP:ro
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
networks:
- vigil_network_dev
ports:
- "${BACKEND_HOST_PORT:-36969}:8000"
restart: unless-stopped
geoip-updater:
image: maxmindinc/geoipupdate:latest
container_name: vigil-geoip-dev
environment:
GEOIPUPDATE_ACCOUNT_ID: ${GEOIP_ACCOUNT_ID:-0}
GEOIPUPDATE_LICENSE_KEY: ${GEOIP_LICENSE_KEY:-placeholder}
GEOIPUPDATE_EDITION_IDS: "GeoLite2-City"
GEOIPUPDATE_FREQUENCY: "168"
volumes:
- geoip_dev_data:/usr/share/GeoIP
networks:
- vigil_network_dev
networks:
vigil_network_dev:
driver: bridge
name: vigil_network_dev
volumes:
postgres_dev_data:
name: vigil_postgres_dev_data
redis_dev_data:
name: vigil_redis_dev_data
geoip_dev_data:
name: vigil_geoip_dev_data

28
PROJECTS/advanced/ai-threat-detection/infra/docker/fastapi.dev Normal file
View File

@ -0,0 +1,28 @@
# ©AngelaMos | 2026
# Development FastAPI Dockerfile
# Hot reload with uvicorn, volume mounts for code
FROM python:3.14-slim
ENV PYTHONUNBUFFERED=1 \
PYTHONDONTWRITEBYTECODE=1
WORKDIR /app
COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/
RUN apt-get update && \
apt-get install -y --no-install-recommends \
build-essential \
libpq-dev \
curl && \
rm -rf /var/lib/apt/lists/*
COPY backend/pyproject.toml ./
RUN uv pip install --system -e .[dev]
COPY backend/ .
EXPOSE 8000
CMD ["uvicorn", "app.factory:create_app", "--factory", "--host", "0.0.0.0", "--port", "8000", "--reload"]

49
PROJECTS/advanced/ai-threat-detection/infra/docker/fastapi.prod Normal file
View File

@ -0,0 +1,49 @@
# ©AngelaMos | 2026
# Production FastAPI Dockerfile
# Multi-stage build, single uvicorn worker (pipeline runs in-process)
FROM python:3.14-slim AS builder
ENV PYTHONUNBUFFERED=1 \
PYTHONDONTWRITEBYTECODE=1
WORKDIR /app
COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/
RUN apt-get update && \
apt-get install -y --no-install-recommends \
build-essential \
libpq-dev && \
rm -rf /var/lib/apt/lists/*
COPY backend/pyproject.toml ./
RUN uv pip install --target /app/deps .
FROM python:3.14-slim
WORKDIR /app
RUN apt-get update && \
apt-get install -y --no-install-recommends \
libpq5 \
curl && \
rm -rf /var/lib/apt/lists/* && \
useradd -m -u 1000 appuser
COPY --from=builder /app/deps /home/appuser/.local/lib/python3.14/site-packages
ENV PYTHONPATH=/home/appuser/.local/lib/python3.14/site-packages:$PYTHONPATH \
PATH=/home/appuser/.local/lib/python3.14/site-packages/bin:$PATH
COPY backend/ .
RUN mkdir -p /app/data/models && \
chown -R appuser:appuser /app
USER appuser
EXPOSE 8000
HEALTHCHECK --interval=30s --timeout=10s --start-period=40s --retries=3 \
CMD curl -f http://localhost:8000/health || exit 1
CMD ["uvicorn", "app.factory:create_app", "--factory", "--host", "0.0.0.0", "--port", "8000"]

58
PROJECTS/advanced/ai-threat-detection/infra/redis/redis.conf Normal file
View File

@ -0,0 +1,58 @@
# ©AngelaMos | 2026
# redis.conf
bind 0.0.0.0
protected-mode yes
rename-command FLUSHALL ""
rename-command FLUSHDB ""
rename-command CONFIG ""
rename-command SHUTDOWN ""
rename-command MONITOR ""
rename-command DEBUG ""
rename-command SLAVEOF ""
rename-command MIGRATE ""
loglevel notice
slowlog-log-slower-than 10000
slowlog-max-len 128
latency-monitor-threshold 100
appendonly yes
appendfsync everysec
save 900 1
save 300 10
save 60 10000
maxmemory 2gb
maxmemory-policy allkeys-lru
maxclients 10000
tracking-table-max-keys 1000000
port 6379
io-threads 4
io-threads-do-reads yes
tcp-backlog 2048
timeout 0
tcp-keepalive 300
activedefrag yes
active-defrag-ignore-bytes 25mb
active-defrag-threshold-lower 5
active-defrag-threshold-upper 20
active-defrag-cycle-min 1
active-defrag-cycle-max 75
lazyfree-lazy-eviction yes
lazyfree-lazy-expire yes
lazyfree-lazy-server-del yes
replica-lazy-flush yes
jemalloc-bg-thread yes

275
PROJECTS/advanced/ai-threat-detection/justfile Normal file
View File

@ -0,0 +1,275 @@
# =============================================================================
# AngelaMos | 2026
# Justfile - AngelusVigil AI Threat Detection
# =============================================================================
set dotenv-load
set export
set shell := ["bash", "-uc"]
set windows-shell := ["powershell.exe", "-NoLogo", "-Command"]
project := file_name(justfile_directory())
version := `git describe --tags --always 2>/dev/null || echo "dev"`
# =============================================================================
# Default
# =============================================================================
default:
@just --list --unsorted
# =============================================================================
# Linting and Formatting
# =============================================================================
[group('lint')]
ruff *ARGS:
cd backend && uv run ruff check app/ cli/ {{ARGS}}
[group('lint')]
ruff-fix:
cd backend && uv run ruff check app/ cli/ --fix
cd backend && uv run ruff format app/ cli/
[group('lint')]
ruff-format:
cd backend && uv run ruff format app/ cli/
[group('lint')]
pylint *ARGS:
cd backend && uv run pylint app {{ARGS}}
[group('lint')]
lint: ruff pylint
# =============================================================================
# Type Checking
# =============================================================================
[group('types')]
mypy *ARGS:
cd backend && uv run mypy app/ cli/ {{ARGS}}
[group('types')]
typecheck: mypy
# =============================================================================
# Testing
# =============================================================================
[group('test')]
pytest *ARGS:
cd backend && uv run pytest tests/ {{ARGS}}
[group('test')]
test: pytest
[group('test')]
test-cov:
cd backend && uv run pytest tests/ --cov=app --cov-report=term-missing --cov-report=html
# =============================================================================
# CI / Quality
# =============================================================================
[group('ci')]
ci: lint typecheck test
[group('ci')]
check: ruff mypy
# =============================================================================
# Docker Compose (Production)
# =============================================================================
[group('docker')]
up *ARGS:
docker compose up {{ARGS}}
[group('docker')]
start *ARGS:
docker compose up -d {{ARGS}}
[group('docker')]
down *ARGS:
docker compose down {{ARGS}}
[group('docker')]
stop:
docker compose stop
[group('docker')]
build *ARGS:
docker compose build {{ARGS}}
[group('docker')]
rebuild:
docker compose build --no-cache
[group('docker')]
logs *SERVICE:
docker compose logs -f {{SERVICE}}
[group('docker')]
ps:
docker compose ps
[group('docker')]
shell service='backend':
docker compose exec -it {{service}} /bin/bash
# =============================================================================
# Docker Compose (Dev)
# =============================================================================
[group('dev')]
dev-up *ARGS:
docker compose -f dev.compose.yml up {{ARGS}}
[group('dev')]
dev-start *ARGS:
docker compose -f dev.compose.yml up -d {{ARGS}}
[group('dev')]
dev-down *ARGS:
docker compose -f dev.compose.yml down {{ARGS}}
[group('dev')]
dev-stop:
docker compose -f dev.compose.yml stop
[group('dev')]
dev-build *ARGS:
docker compose -f dev.compose.yml build {{ARGS}}
[group('dev')]
dev-rebuild:
docker compose -f dev.compose.yml build --no-cache
[group('dev')]
dev-logs *SERVICE:
docker compose -f dev.compose.yml logs -f {{SERVICE}}
[group('dev')]
dev-ps:
docker compose -f dev.compose.yml ps
[group('dev')]
dev-shell service='backend':
docker compose -f dev.compose.yml exec -it {{service}} /bin/bash
# =============================================================================
# Database (Docker)
# =============================================================================
[group('db')]
migrate *ARGS:
docker compose exec backend uv run alembic upgrade {{ARGS}}
[group('db')]
migration message:
docker compose exec backend uv run alembic revision --autogenerate -m "{{message}}"
[group('db')]
rollback:
docker compose exec backend uv run alembic downgrade -1
[group('db')]
db-history:
docker compose exec backend uv run alembic history --verbose
[group('db')]
db-current:
docker compose exec backend uv run alembic current
# =============================================================================
# Database (Local - no Docker)
# =============================================================================
[group('db-local')]
migrate-local *ARGS:
cd backend && uv run alembic upgrade {{ARGS}}
[group('db-local')]
migration-local message:
cd backend && uv run alembic revision --autogenerate -m "{{message}}"
[group('db-local')]
rollback-local:
cd backend && uv run alembic downgrade -1
[group('db-local')]
db-history-local:
cd backend && uv run alembic history --verbose
[group('db-local')]
db-current-local:
cd backend && uv run alembic current
# =============================================================================
# AngelusVigil CLI
# =============================================================================
[group('vigil')]
serve *ARGS:
cd backend && uv run python -m app {{ARGS}}
[group('vigil')]
vigil-config:
cd backend && uv run vigil config
[group('vigil')]
vigil-health:
curl -sf http://localhost:8000/health | python -m json.tool
# =============================================================================
# Setup
# =============================================================================
[group('setup')]
setup: setup-backend env
@echo "Setup complete!"
[group('setup')]
setup-backend:
@echo "Setting up backend..."
cd backend && uv sync
@echo "Backend setup complete!"
[group('setup')]
env:
@echo "Creating .env file..."
@if [ ! -f .env ]; then cp .env.example .env; echo "Created .env from example"; fi
@echo ".env ready! Please update with your values."
# =============================================================================
# Utilities
# =============================================================================
[group('util')]
info:
@echo "Project: {{project}}"
@echo "Version: {{version}}"
@echo "OS: {{os()}} ({{arch()}})"
[group('util')]
clean:
-rm -rf backend/.mypy_cache
-rm -rf backend/.pytest_cache
-rm -rf backend/.ruff_cache
-rm -rf backend/htmlcov
-rm -rf backend/.coverage
@echo "Cache directories cleaned"
[group('util')]
[confirm("Remove all containers, volumes, and build artifacts?")]
nuke:
@echo "Nuking everything..."
-docker compose -f dev.compose.yml down -v
-docker compose down -v
-rm -rf backend/.venv
-rm -rf backend/__pycache__
-rm -rf backend/.pytest_cache
-rm -rf backend/.mypy_cache
-rm -rf backend/.ruff_cache
@echo "Nuke complete!"