@guanlin4924/devflow-pi 0.1.77 → 0.1.78

package/plugins/devflow-intel/skills/workflow-code-index/scripts/writers/sqlite_writer.py

@@ -2,6 +2,7 @@ import sqlite3
 import os
 from pathlib import Path
 
+from adapters.capabilities import confidence_for_parser
 from core.schema import json_dumps
 
 
@@ -31,6 +32,7 @@ CREATE TABLE IF NOT EXISTS symbols (
   language TEXT NOT NULL,
   content_hash TEXT NOT NULL,
   visibility TEXT,
+  parser TEXT NOT NULL,
   metadata_json TEXT
 );
 
@@ -113,6 +115,7 @@ CREATE TABLE IF NOT EXISTS feature_bindings (
   confidence TEXT NOT NULL,
   source_path TEXT,
   line_start INTEGER,
+  extractor TEXT NOT NULL,
   PRIMARY KEY(snapshot_id, feature_id, entity_id)
 );
 
@@ -207,19 +210,27 @@ def write_sqlite(path: Path, files, symbols, relations, diagnostics, snapshot_id
         parsed_count = sum(1 for item in files if item.get("path") not in parse_failed_paths)
         conn.execute(
             "INSERT INTO analysis_snapshots(snapshot_id, schema_version, source_revision, content_manifest_hash, extractor_set_hash, created_at, status) VALUES(?,?,?,?,?,?,?)",
-            (snapshot_id, 2, workspace_state.get("repositoryRevision", ""), workspace_state.get("contentManifestHash", ""),
+            (snapshot_id, 3, workspace_state.get("repositoryRevision", ""), workspace_state.get("contentManifestHash", ""),
              workspace_state.get("extractorSetHash", ""), workspace_state.get("generatedAt", ""), "current"),
         )
         # Structural symbols and files are first-class graph nodes as well.
         # This prevents semantic flows and feature bindings from dangling when
         # a framework linker points at an existing code symbol.
-        structural_entities = [
-            {"id": item["id"], "entity_type": item.get("kind", "symbol"), "name": item.get("name", item["id"]),
-             "codebase": item.get("codebase"), "path": item.get("path"), "line_start": item.get("line_start"),
-             "line_end": item.get("line_end"), "content_hash": item.get("content_hash"), "confidence": "exact",
-             "extractor": item.get("parser", "index"), "metadata": item.get("metadata", {})}
-            for item in [*files, *symbols]
-        ]
+        parser_by_path = {item.get("path"): item.get("parser", "unknown") for item in files}
+        structural_entities = []
+        for item in [*files, *symbols]:
+            is_file = "size_bytes" in item
+            metadata = dict(item.get("metadata", {}))
+            parser = item.get("parser") or metadata.get("parser") or parser_by_path.get(item.get("path"), "unknown")
+            metadata.setdefault("parser", parser)
+            structural_entities.append({
+                "id": item["id"], "entity_type": item.get("kind", "file" if is_file else "symbol"),
+                "name": item.get("name", item.get("path", item["id"])), "codebase": item.get("codebase"),
+                "path": item.get("path"), "line_start": item.get("line_start", 1 if is_file else None),
+                "line_end": item.get("line_end", 1 if is_file else None), "content_hash": item.get("content_hash"),
+                "confidence": confidence_for_parser(parser, "file" if is_file else "symbol"),
+                "extractor": parser, "metadata": metadata,
+            })
         by_id = {item["id"]: item for item in semantic_entities}
         for item in structural_entities:
             by_id.setdefault(item["id"], item)
@@ -247,8 +258,8 @@ def write_sqlite(path: Path, files, symbols, relations, diagnostics, snapshot_id
             [(snapshot_id, item["source_id"], item["target_id"], item["flow_type"], item.get("source_path"), item.get("line_start"), item["confidence"], item.get("evidence_level", "static_source"), item.get("extractor", "unknown"), json_dumps(item.get("metadata", {}))) for item in semantic_flows],
         )
         conn.executemany(
-            "INSERT INTO feature_bindings(snapshot_id, feature_id, entity_id, binding_type, confidence, source_path, line_start) VALUES(?,?,?,?,?,?,?)",
-            [(snapshot_id, item["feature_id"], item["entity_id"], item["binding_type"], item["confidence"], item.get("source_path"), item.get("line_start")) for item in feature_bindings],
+            "INSERT INTO feature_bindings(snapshot_id, feature_id, entity_id, binding_type, confidence, source_path, line_start, extractor) VALUES(?,?,?,?,?,?,?,?)",
+            [(snapshot_id, item["feature_id"], item["entity_id"], item["binding_type"], item["confidence"], item.get("source_path"), item.get("line_start"), item.get("extractor", "unknown")) for item in feature_bindings],
         )
         conn.execute(
             "INSERT INTO coverage_metrics(snapshot_id, files_scanned, files_parsed, diagnostics, semantic_entities, semantic_flows, dynamic_limitations, metadata_json) VALUES(?,?,?,?,?,?,?,?)",
@@ -264,9 +275,9 @@ def write_sqlite(path: Path, files, symbols, relations, diagnostics, snapshot_id
         conn.executemany(
             """
             INSERT INTO symbols(id, codebase, kind, name, qualified_name, container_id, path, line_start, line_end,
-              signature, language, content_hash, visibility, metadata_json)
+              signature, language, content_hash, visibility, parser, metadata_json)
             VALUES(:id, :codebase, :kind, :name, :qualified_name, :container_id, :path, :line_start, :line_end,
-              :signature, :language, :content_hash, :visibility, :metadata_json)
+              :signature, :language, :content_hash, :visibility, :parser, :metadata_json)
             """,
             [_symbol_row(record) for record in symbols],
         )
@@ -302,7 +313,9 @@ def write_sqlite(path: Path, files, symbols, relations, diagnostics, snapshot_id
 
 def _symbol_row(record):
     row = dict(record)
-    row["metadata_json"] = json_dumps(row.pop("metadata", {}))
+    metadata = row.pop("metadata", {})
+    row["parser"] = row.get("parser") or metadata.get("parser") or "unknown"
+    row["metadata_json"] = json_dumps(metadata)
     return row
 
 

package/src/constants.ts

@@ -1,5 +1,5 @@
 export const PRODUCT_NAME = "Devflow Pi";
-export const PRODUCT_VERSION = "0.1.77";
+export const PRODUCT_VERSION = "0.1.78";
 export const PROJECT_SCHEMA_VERSION = 1;
 export const OMP_RUNTIME_VERSION = "17.2.7";