foldkeep 0.9.2__tar.gz → 0.9.3__tar.gz

{foldkeep-0.9.2 → foldkeep-0.9.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: foldkeep
-Version: 0.9.2
+Version: 0.9.3
 Summary: Zero-dependency lossless context folding for agent sessions: pinned facts, tolerance routing, byte-for-byte expand.
 Author: foldkeep contributors
 License: MIT License
@@ -37,6 +37,9 @@ Dynamic: license-file
 
 # foldkeep
 
+[![PyPI](https://img.shields.io/pypi/v/foldkeep)](https://pypi.org/project/foldkeep/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
+
 **Fold the context. Keep the facts.**
 
 foldkeep is a zero-dependency, deterministic context-compression engine for
@@ -46,6 +49,12 @@ you expand ANY turn back byte-for-byte, even after rolling and merging.
 
 Pure Python stdlib. No LLM calls. Millisecond latency. SQLite-backed.
 
+## Install
+
+```bash
+pip install foldkeep
+```
+
 ## Why
 
 Agent transcripts grow until the model drowns. Naive fixes lose the past:

{foldkeep-0.9.2 → foldkeep-0.9.3}/README.md

@@ -1,5 +1,8 @@
 # foldkeep
 
+[![PyPI](https://img.shields.io/pypi/v/foldkeep)](https://pypi.org/project/foldkeep/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
+
 **Fold the context. Keep the facts.**
 
 foldkeep is a zero-dependency, deterministic context-compression engine for
@@ -9,6 +12,12 @@ you expand ANY turn back byte-for-byte, even after rolling and merging.
 
 Pure Python stdlib. No LLM calls. Millisecond latency. SQLite-backed.
 
+## Install
+
+```bash
+pip install foldkeep
+```
+
 ## Why
 
 Agent transcripts grow until the model drowns. Naive fixes lose the past:
@@ -77,4 +86,4 @@ v0.9.2. Research notes (`RESEARCH.md`) document the competitive landscape
 (LLMLingua, leanctx, mnesis, FoldAgent, pi-fold, ECC, hermes-agent) and what
 we learned from each.
 
-MIT licensed. Windows/macOS/Linux, Python 3.10+.
+MIT licensed. Windows/macOS/Linux, Python 3.10+.

{foldkeep-0.9.2 → foldkeep-0.9.3}/foldkeep/session.py

@@ -91,6 +91,19 @@ def _conn():
     c.execute("""CREATE TABLE IF NOT EXISTS originals (
         name TEXT NOT NULL, seq INTEGER NOT NULL,
         text TEXT NOT NULL, PRIMARY KEY (name, seq))""")
+    # v0.9.3: entity frequency/recency stats (ECC instinct-style ranking)
+    c.execute("""CREATE TABLE IF NOT EXISTS entity_stats (
+        name TEXT NOT NULL, ent TEXT NOT NULL,
+        freq INTEGER NOT NULL DEFAULT 0,
+        last_seq INTEGER NOT NULL DEFAULT 0,
+        PRIMARY KEY (name, ent))""")
+    # v0.9.3: FTS5 mirror of originals (hermes-style session search);
+    # originals are immutable so the index never needs updating
+    try:
+        c.execute("CREATE VIRTUAL TABLE IF NOT EXISTS search_fts "
+                  "USING fts5(name, seq, text)")
+    except sqlite3.OperationalError:
+        pass    # FTS5 unavailable -> search() falls back to LIKE
     return c
 
 _FENCE_RE = re.compile(r"```|~~~")
@@ -120,6 +133,12 @@ def new(name, project="default"):
         c.execute("DELETE FROM turns WHERE name=?", (name,))
         c.execute("DELETE FROM entities WHERE name=?", (name,))
         c.execute("DELETE FROM facts WHERE name=?", (name,))
+        c.execute("DELETE FROM entity_stats WHERE name=?", (name,))
+        c.execute("DELETE FROM originals WHERE name=?", (name,))
+        try:
+            c.execute("DELETE FROM search_fts WHERE name=?", (name,))
+        except sqlite3.OperationalError:
+            pass
     return name
 
 def _rows(name):
@@ -127,7 +146,9 @@ def _rows(name):
         return c.execute("SELECT sid,seq,role,kind,text,orig_chars FROM turns "
                          "WHERE name=? ORDER BY seq", (name,)).fetchall()
 
-def push(name, role, text, roll_threshold=2600):
+def push(name, role, text, roll_threshold=None):
+    if roll_threshold is None:
+        roll_threshold = _env_int("OC_ROLL_THRESHOLD", 2600)
     with _conn() as c:
         mx = c.execute("SELECT COALESCE(MAX(seq),0) FROM turns WHERE name=?",
                        (name,)).fetchone()[0]
@@ -135,6 +156,11 @@ def push(name, role, text, roll_threshold=2600):
                   "created,orig_text) VALUES (?,?,?,?,?,?,?,?)",
                   (name, mx + 1, role, "raw", text, len(text), time.time(),
                    text))
+        try:
+            c.execute("INSERT INTO search_fts (name, seq, text) "
+                      "VALUES (?,?,?)", (name, mx + 1, text))
+        except sqlite3.OperationalError:
+            pass
         # pin a LEAN fact line: context windows around entities/numbers
         # of the salient sentence -- exact substrings survive, filler
         # does not; render shows it only for out-of-body turns
@@ -189,6 +215,21 @@ def _rollup(name, threshold):
         # entity registry: rolled-away turns keep entity facts reachable
         for e in ents:
             c.execute("INSERT OR IGNORE INTO entities VALUES (?,?)", (name, e))
+        # v0.9.3 frequency/recency stats for relevance-ranked injection
+        from collections import Counter
+        cnt, last = Counter(), {}
+        for r in rows[:cut]:
+            if _intolerant(r[4]):
+                continue
+            for e in set(ENT_RE.findall(r[4])):
+                cnt[e] += 1
+                last[e] = max(last.get(e, 0), r[1])
+        for e, f in cnt.items():
+            c.execute("INSERT INTO entity_stats VALUES (?,?,?,?) "
+                      "ON CONFLICT(name, ent) DO UPDATE SET "
+                      "freq=freq+excluded.freq, "
+                      "last_seq=MAX(last_seq, excluded.last_seq)",
+                      (name, e, f, last[e]))
     _consolidate(name)
     return n
 
@@ -235,6 +276,20 @@ def _registry(name, cap=None):
     if cap is None:
         cap = _env_int("OC_REGISTRY_MAX", 48)
     with _conn() as c:
+        # v0.9.3 relevance ranking (ECC instinct-style), TIERED so the
+        # old SCREAMING-first guarantee still holds: SCREAMING_CASE
+        # entities always outrank lowercase ones; within a tier,
+        # frequency x recency decides, alphabetical tiebreak.
+        # (Pure freq ranking regressed the marathon: high-freq lowercase
+        # filler evicted low-freq SCREAMING checkpoints from the cap.)
+        rows = c.execute(
+            "SELECT ent, freq, last_seq FROM entity_stats WHERE name=?"
+            , (name,)).fetchall()
+        if rows:
+            rows.sort(key=lambda r: (
+                0 if re.match(r"^[A-Z][A-Z0-9_]{2,}$", r[0]) else 1,
+                -r[1], -r[2], r[0]))
+            return [r[0] for r in rows][:cap]
         rows = c.execute("SELECT DISTINCT ent FROM entities WHERE name=?",
                          (name,)).fetchall()
     ents = [r[0] for r in rows]
@@ -371,15 +426,32 @@ def search(name, query, limit=10):
     search): substring match over verbatim originals + pinned facts,
     newest first. Works even for turns that were rolled away or merged,
     because originals survive (v0.9 lossless base)."""
-    pat = "%" + query.replace("%", "").replace("_", " ") + "%"
     like = "%" + query.replace("%", "") + "%"
     with _conn() as c:
-        rows = c.execute(
-            "SELECT seq, role, COALESCE(orig_text, text) FROM turns "
-            "WHERE name=? AND (COALESCE(orig_text, text) LIKE ? "
-            "OR COALESCE(orig_text, text) LIKE ?) "
-            "ORDER BY seq DESC LIMIT ?",
-            (name, like, pat, limit)).fetchall()
+        rows = None
+        try:
+            # v0.9.3 FTS5 over the originals mirror (hermes-style);
+            # every term ANDed, quoted to avoid syntax issues
+            match = " ".join('"%s"' % t.replace('"', "")
+                             for t in query.split() if t)
+            if match:
+                rows = c.execute(
+                    "SELECT f.seq, COALESCE(t.role,'system'), f.text "
+                    "FROM search_fts f LEFT JOIN turns t "
+                    "ON t.name=f.name AND t.seq=f.seq "
+                    "WHERE search_fts MATCH ? AND f.name=? "
+                    "ORDER BY f.seq DESC LIMIT ?",
+                    (match, name, limit)).fetchall()
+        except sqlite3.OperationalError:
+            rows = None
+        if rows is None:
+            pat = "%" + query.replace("%", "").replace("_", " ") + "%"
+            rows = c.execute(
+                "SELECT seq, role, COALESCE(orig_text, text) FROM turns "
+                "WHERE name=? AND (COALESCE(orig_text, text) LIKE ? "
+                "OR COALESCE(orig_text, text) LIKE ?) "
+                "ORDER BY seq DESC LIMIT ?",
+                (name, like, pat, limit)).fetchall()
         facts = c.execute(
             "SELECT seq, text FROM facts WHERE name=? AND text LIKE ? "
             "ORDER BY seq DESC LIMIT ?", (name, like, limit)).fetchall()

{foldkeep-0.9.2 → foldkeep-0.9.3}/foldkeep.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: foldkeep
-Version: 0.9.2
+Version: 0.9.3
 Summary: Zero-dependency lossless context folding for agent sessions: pinned facts, tolerance routing, byte-for-byte expand.
 Author: foldkeep contributors
 License: MIT License
@@ -37,6 +37,9 @@ Dynamic: license-file
 
 # foldkeep
 
+[![PyPI](https://img.shields.io/pypi/v/foldkeep)](https://pypi.org/project/foldkeep/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
+
 **Fold the context. Keep the facts.**
 
 foldkeep is a zero-dependency, deterministic context-compression engine for
@@ -46,6 +49,12 @@ you expand ANY turn back byte-for-byte, even after rolling and merging.
 
 Pure Python stdlib. No LLM calls. Millisecond latency. SQLite-backed.
 
+## Install
+
+```bash
+pip install foldkeep
+```
+
 ## Why
 
 Agent transcripts grow until the model drowns. Naive fixes lose the past:

{foldkeep-0.9.2 → foldkeep-0.9.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "foldkeep"
-version = "0.9.2"
+version = "0.9.3"
 description = "Zero-dependency lossless context folding for agent sessions: pinned facts, tolerance routing, byte-for-byte expand."
 readme = "README.md"
 license = { file = "LICENSE" }