From ae840beaa325a790edcd5be340697f85d44e9957 Mon Sep 17 00:00:00 2001
From: "fatih.bulut" <bulutfatih@gmail.com>
Date: Mon, 22 Jun 2026 00:52:45 +0300
Subject: [PATCH] feat(market-research): add history.py non-repeat memory with
 offline test

---
 .../skills/market-research/scripts/history.py | 83 +++++++++++++++++++
 .../tests/fixtures/candidates-sample.json     |  5 ++
 .../tests/fixtures/history-seed.json          |  6 ++
 plugins/market-research/tests/test-history.py | 51 ++++++++++++
 4 files changed, 145 insertions(+)
 create mode 100755 plugins/market-research/skills/market-research/scripts/history.py
 create mode 100644 plugins/market-research/tests/fixtures/candidates-sample.json
 create mode 100644 plugins/market-research/tests/fixtures/history-seed.json
 create mode 100644 plugins/market-research/tests/test-history.py

diff --git a/plugins/market-research/skills/market-research/scripts/history.py b/plugins/market-research/skills/market-research/scripts/history.py
new file mode 100755
index 0000000..16ba9b9
--- /dev/null
+++ b/plugins/market-research/skills/market-research/scripts/history.py
@@ -0,0 +1,83 @@
+#!/usr/bin/env python3
+"""Non-repeat memory for market-research suggestions.
+
+history.json holds every candidate ever suggested so a run can avoid repeating
+itself. Identity key = package if present, else the lowercased/stripped name.
+
+Subcommands (candidates read as a JSON array on stdin):
+  filter --history H.json   -> print the candidates minus already-seen ones
+  add    --history H.json   -> append candidates to H.json, print {added,total}
+
+Date/run_id for added entries come from --date / --run-id (the skill passes the
+run date); both default to empty strings so the script is deterministic and
+needs no clock. Stdlib-only, no pip.
+"""
+import sys, json, argparse, os
+
+def cand_key(entry):
+    """Canonical identity for a candidate. Package wins; else lowercased name."""
+    pkg = entry.get("package")
+    if pkg:
+        return str(pkg).strip()
+    return str(entry.get("name", "")).strip().lower()
+
+def load_history(path):
+    if not os.path.exists(path):
+        return {"suggestions": []}
+    with open(path, encoding="utf-8") as f:
+        data = json.load(f)
+    data.setdefault("suggestions", [])
+    return data
+
+def seen_keys(history):
+    return {s.get("key") for s in history["suggestions"]}
+
+def read_candidates():
+    data = json.load(sys.stdin)
+    if not isinstance(data, list):
+        raise ValueError("candidates stdin must be a JSON array")
+    return data
+
+def cmd_filter(args):
+    history = load_history(args.history)
+    seen = seen_keys(history)
+    cands = read_candidates()
+    survivors = [c for c in cands if cand_key(c) not in seen]
+    print(json.dumps(survivors, indent=2))
+
+def cmd_add(args):
+    history = load_history(args.history)
+    seen = seen_keys(history)
+    cands = read_candidates()
+    added = 0
+    for c in cands:
+        k = cand_key(c)
+        if k in seen:
+            continue
+        history["suggestions"].append({
+            "key": k,
+            "name": c.get("name"),
+            "package": c.get("package"),
+            "date": args.date,
+            "run_id": args.run_id,
+        })
+        seen.add(k)
+        added += 1
+    os.makedirs(os.path.dirname(os.path.abspath(args.history)), exist_ok=True)
+    with open(args.history, "w", encoding="utf-8") as f:
+        json.dump(history, f, indent=2)
+    print(json.dumps({"added": added, "total": len(history["suggestions"])}, indent=2))
+
+def main():
+    ap = argparse.ArgumentParser()
+    sub = ap.add_subparsers(dest="cmd", required=True)
+    for name in ("filter", "add"):
+        p = sub.add_parser(name)
+        p.add_argument("--history", required=True)
+        p.add_argument("--date", default="")
+        p.add_argument("--run-id", default="")
+    args = ap.parse_args()
+    {"filter": cmd_filter, "add": cmd_add}[args.cmd](args)
+
+if __name__ == "__main__":
+    main()
diff --git a/plugins/market-research/tests/fixtures/candidates-sample.json b/plugins/market-research/tests/fixtures/candidates-sample.json
new file mode 100644
index 0000000..9e2af50
--- /dev/null
+++ b/plugins/market-research/tests/fixtures/candidates-sample.json
@@ -0,0 +1,5 @@
+[
+  { "name": "Puzzle Quest Saga", "package": "com.casual.puzzlequest", "category": "Games" },
+  { "name": "habit tracker PRO", "category": "Productivity" },
+  { "name": "Budget Buddy", "package": "com.fintech.budgetbuddy", "category": "Finance" }
+]
diff --git a/plugins/market-research/tests/fixtures/history-seed.json b/plugins/market-research/tests/fixtures/history-seed.json
new file mode 100644
index 0000000..13db63f
--- /dev/null
+++ b/plugins/market-research/tests/fixtures/history-seed.json
@@ -0,0 +1,6 @@
+{
+  "suggestions": [
+    { "key": "com.casual.puzzlequest", "name": "Puzzle Quest Saga", "package": "com.casual.puzzlequest", "date": "2026-06-01", "run_id": "seed" },
+    { "key": "habit tracker pro", "name": "Habit Tracker Pro", "package": null, "date": "2026-06-01", "run_id": "seed" }
+  ]
+}
diff --git a/plugins/market-research/tests/test-history.py b/plugins/market-research/tests/test-history.py
new file mode 100644
index 0000000..63ad544
--- /dev/null
+++ b/plugins/market-research/tests/test-history.py
@@ -0,0 +1,51 @@
+#!/usr/bin/env python3
+import json, subprocess, sys, os, tempfile, shutil
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+SCRIPT = os.path.join(HERE, "..", "skills", "market-research", "scripts", "history.py")
+FIX = os.path.join(HERE, "fixtures")
+
+def run(args, stdin_path):
+    with open(stdin_path, "rb") as f:
+        return subprocess.run([sys.executable, SCRIPT, *args],
+                              stdin=f, capture_output=True)
+
+def main():
+    fails = []
+    def check(name, cond):
+        print(f"{'PASS' if cond else 'FAIL'}: {name}")
+        if not cond: fails.append(name)
+
+    cands = os.path.join(FIX, "candidates-sample.json")
+
+    # filter: against the seed, only "Budget Buddy" survives
+    r = run(["filter", "--history", os.path.join(FIX, "history-seed.json")], cands)
+    check("filter exit 0", r.returncode == 0)
+    survivors = json.loads(r.stdout)
+    names = sorted(c["name"] for c in survivors)
+    check("filter drops package collision + name collision", names == ["Budget Buddy"])
+
+    # filter against a missing history file = nothing seen, all 3 survive
+    tmp = tempfile.mkdtemp()
+    try:
+        missing = os.path.join(tmp, "nope.json")
+        r = run(["filter", "--history", missing], cands)
+        check("filter missing-history exit 0", r.returncode == 0)
+        check("filter missing-history keeps all", len(json.loads(r.stdout)) == 3)
+
+        # add: appends all 3 to a fresh history, reports counts
+        h = os.path.join(tmp, "h.json")
+        r = run(["add", "--history", h], cands)
+        check("add exit 0", r.returncode == 0)
+        rep = json.loads(r.stdout)
+        check("add reports added 3", rep["added"] == 3)
+        check("add reports total 3", rep["total"] == 3)
+        saved = json.load(open(h))
+        check("history file has 3 suggestions", len(saved["suggestions"]) == 3)
+        check("history entries carry key", all("key" in s for s in saved["suggestions"]))
+    finally:
+        shutil.rmtree(tmp)
+
+    sys.exit(1 if fails else 0)
+
+main()