7aeebc45ea
Merge 551b084021 into b5b53acc57
2026-08-09 03:26:22 +00:00
551b084021
fix: respect enabled platforms when checking preparation
2026-08-09 00:08:11 +07:00
b5b53acc57
chore: update Star History ( #755 )
...
Automated aggregate Star History refresh. Recorded 69,857 stars.
2026-08-03 18:19:13 +08:00
0f67b29c6c
fix: submit Star History updates as pull requests ( #754 )
...
The workflow now creates and verifies a pull request for manual owner merge without attempting to update main itself.
2026-08-03 18:12:17 +08:00
a97ba4ffc9
Fix automated Star History updates on protected main ( #752 )
...
Replace direct pushes with a validated temporary-branch and pull-request publication flow.
2026-08-03 14:24:45 +08:00
60757b3c82
Merge pull request #742 from 666ghj/codex/fix-ontology-provider-compatibility
...
fix: harden ontology JSON generation across LLM providers
2026-07-23 12:47:16 +08:00
f0d487f310
fix: harden ontology JSON generation
2026-07-23 02:08:17 +08:00
d64175943e
Merge pull request #740 from 666ghj/codex/zep-cloud-comprehensive-audit
...
fix: modernize Zep Cloud integration and dependency security
2026-07-23 00:44:26 +08:00
407c391f35
fix: keep Zep timeouts internal
2026-07-23 00:32:53 +08:00
c81d628975
chore: update vulnerable npm dependencies
2026-07-23 00:26:51 +08:00
54bd7a3d4e
fix: validate complete Zep entity context
2026-07-23 00:20:50 +08:00
e4aa38dff2
fix: modernize Zep Cloud integration
2026-07-23 00:20:28 +08:00
dd7a646d6b
Merge pull request #739 : use the supported Zep node-edge method
...
Call graph.node.get_edges so entity enrichment no longer retries an absent SDK method and returns empty edges.
2026-07-22 20:08:45 +08:00
0d8c23a3c2
fix: use the supported Zep node-edge method
2026-07-22 20:08:18 +08:00
aa6abf3e39
Merge pull request #738 : normalize ontology fields for Zep
...
Normalize untrusted LLM ontology fields and enforce the pinned Zep schema limits before upload.
2026-07-22 19:54:36 +08:00
d2f9e56eea
fix: enforce Zep ontology contract limits
2026-07-22 19:53:59 +08:00
9506e6f5e0
fix: normalize ontology attributes centrally
2026-07-22 19:51:41 +08:00
a07841e91b
fix: handle string attributes in graph ontology to prevent TypeError crash ( fixes #135 )
...
When the LLM returns ontology attributes as plain strings instead of
dicts, set_ontology() crashes with "TypeError: string indices must be
integers, not 'str'" at attr_def["name"].
Two-layer fix:
1. ontology_generator.py: normalize string attrs to {"name", "type",
"description"} dicts during validation, so downstream code always
receives well-formed structures.
2. graph_builder.py: add isinstance guard as a safety net in
set_ontology() for both entity and edge attribute loops.
Co-Authored-By: Octopus <liyuan851277048@icloud.com>
2026-07-22 19:51:19 +08:00
4064b68abf
Merge pull request #737 : normalize generated profile fields
...
Establish one canonical profile boundary before Twitter and Reddit serialization.
2026-07-22 19:51:02 +08:00
c415387995
style: remove trailing whitespace from profile model
2026-07-22 19:50:18 +08:00
9f80ad66fd
fix: normalize profile fields at construction
2026-07-22 19:49:33 +08:00
41fc932b62
fix: normalize structured LLM profile fields before serialization
2026-07-22 19:49:33 +08:00
666a188f6b
Merge pull request #736 : handle GPT-5 parameters deterministically
...
Use explicit GPT-5 Chat Completions parameters while preserving legacy requests and provider errors.
2026-07-22 19:49:18 +08:00
e67a52e80f
fix: keep GPT-5 compatibility handling narrow
2026-07-22 19:48:32 +08:00
f7138e09fb
Correction GPT-5 family error 500
2026-07-22 19:48:32 +08:00
2576c06f95
Merge pull request #735 : stop Step 2 after preparation failure
...
Propagate terminal preparation failure and stop all Step 2 polling without fabricating Zep entity labels.
2026-07-22 19:48:14 +08:00
05bea4a330
fix: stop Step2 polling after preparation failure
2026-07-22 19:47:20 +08:00
7970d74a86
Merge pull request #734 : sanitize fabricated tool results
...
Remove fabricated tool results consistently before assistant content enters trusted history.
2026-07-22 19:47:02 +08:00
f0e379471a
fix: sanitize fabricated tool results everywhere
2026-07-22 19:46:23 +08:00
1d96b09875
fix: strip fabricated tool_result blocks to prevent LLM hallucination
...
When the LLM generates a <tool_call> block followed by a self-generated
<tool_result> block in the same response, the fake result must be stripped
before appending to message history. The real tool result will be injected
separately by the system.
This fixes a React Hallucination bug where models could fabricate tool
results that didn't actually come from tool invocations.
2026-07-22 19:46:23 +08:00
ca845b0fcd
Merge pull request #733 : support Twitter-only simulation retrieval
...
Select the configured platform and load Twitter profiles from their actual CSV output.
2026-07-22 19:46:08 +08:00
b651883d66
fix: load Twitter profiles from CSV
2026-07-22 19:45:30 +08:00
aaf3ea0a09
Fix silent data loss when platform defaults to reddit for Twitter-only simulations
...
The API retrieval layer hardcoded reddit as the default platform in 11+ locations. When a Twitter-only simulation was run, all data retrieval APIs silently returned empty results because they looked for reddit_simulation.db which did not exist. This commit reads the simulation enable_twitter/enable_reddit config to determine the correct default platform.
Fixes #150
2026-07-22 19:45:30 +08:00
e16456575e
Merge pull request #732 : prioritize terminal simulation failure
...
Stop polling on explicit runner failure even when platform counters could imply completion.
2026-07-22 19:45:17 +08:00
8b414607c7
fix: prioritize failed simulation status
2026-07-22 19:44:42 +08:00
867b54cb81
fix: stop polling when simulation fails
...
Previously, the frontend only checked for 'completed' and 'stopped'
status to stop polling. When the simulation process exited with an
error (runner_status === 'failed'), polling continued indefinitely.
This commit adds handling for the 'failed' status:
- Stop polling when runner_status is 'failed'
- Display error message in the log
- Update UI status to 'error'
- Add i18n translations for simFailed key
2026-07-22 19:44:42 +08:00
66f2bb6e70
Merge pull request #731 : add guarded ARM64 Docker publishing
...
Build amd64 and arm64 images while making manual publication explicit and default-branch-only.
2026-07-22 19:44:11 +08:00
898f3b5bea
ci: make manual Docker publishing opt-in
2026-07-22 19:43:35 +08:00
564a5603d5
ci: build docker image for arm64
2026-07-22 19:43:35 +08:00
ce616b05f1
Merge pull request #730 : cap Zep edge pagination safely
...
Add a bounded edge paginator while preserving the existing positional retry API.
2026-07-22 19:43:11 +08:00
7d6f6bb135
fix: preserve edge paging call compatibility
2026-07-22 19:42:00 +08:00
e58d4f1cc5
fix: add max_items limit to fetch_all_edges to prevent unbounded memory growth ( #513 )
...
fetch_all_nodes already had a max_items guard (default 2000) but
fetch_all_edges had no such safeguard, allowing unbounded memory growth
on graphs with large numbers of edges.
Add _MAX_EDGES = 5000 constant and mirror the same loop-guard pattern
from fetch_all_nodes: cap the result list, emit a warning log, and
break pagination once the limit is reached.
2026-07-22 19:42:00 +08:00
53af276f4e
Merge pull request #584 : sample long ontology input across the document
...
Replace first-50000-character truncation with deterministic sampling under the same context budget.
2026-07-22 19:40:54 +08:00
ad12d412ee
Merge pull request #598 : preserve repeated user messages in chat history
...
Remove only the current trailing message instead of filtering every earlier message with equal content.
2026-07-22 19:40:49 +08:00
fbdae298d0
Merge pull request #720 : keep history date and time in the same local timezone
...
Derive both displayed values from local date components so the card cannot show mismatched days.
2026-07-22 19:40:44 +08:00
231ec22bea
Merge pull request #445 : disable Flask debug mode by default
...
Default Flask debug mode to false while preserving explicit opt-in through FLASK_DEBUG.
2026-07-22 19:40:39 +08:00
cbfd03db69
Merge pull request #727 : upgrade NLTK to 3.10.0
...
Upgrade the transitive Python dependency NLTK from 3.9.2 to 3.10.0 to address CVE-2025-14009. The contributor title said 3.9.3, but the reviewed lockfile resolves 3.10.0.
2026-07-22 19:40:34 +08:00
086f1fb5e8
Merge pull request #305 : fix graph tasks endpoint crash
...
Remove the redundant to_dict call because TaskManager.list_tasks already returns dictionaries.
2026-07-22 19:39:57 +08:00
fe2fd5b164
fix: CVE-2025-14009 security vulnerability
...
Automated dependency upgrade by OrbisAI Security
2026-07-21 09:08:09 +00:00
4ebb2ed6e9
feat: localize star history
2026-07-21 15:46:53 +08:00
NgoQuocViet2001
NgoQuocViet2001
github-actions[bot]
BaiFu
octo-patch
ygh1254
IagoMelo01
MiroFish Contributor
karesansui
cuixiao
Sami Rusani
harshitanand
orbisai0security