first commit

src/Vector/VectorSearchClient.php

@@ -0,0 +1,55 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Vector;
+
+use Psr\Log\LoggerInterface;
+
+final class VectorSearchClient
+{
+    public function __construct(
+        private readonly string $vectorDir,
+        private LoggerInterface $agentLogger,
+    ) {
+    }
+
+    public function search(string $query, int $limit = 5): array
+    {
+        $script = rtrim($this->vectorDir, '/') . '/vector_search.py';
+        $this->agentLogger->info("Run vector search script $script");
+        if (!is_file($script)) {
+            return [];
+        }
+
+        // -------------------------------------------------
+        // Determine Python interpreter (venv preferred)
+        // -------------------------------------------------
+        $venvPython = $this->vectorDir . '/.venv/bin/python';
+        $pythonBin  = is_file($venvPython) ? $venvPython : 'python3';
+
+        $cmd = sprintf(
+            '%s %s %s %d 2>&1',
+            escapeshellarg($pythonBin),
+            escapeshellarg($script),
+            escapeshellarg($query),
+            $limit
+        );
+
+        exec($cmd, $out, $exitCode);
+
+        if ($exitCode !== 0 || empty($out)) {
+            return [];
+        }
+
+        $json = implode("\n", $out);
+
+        $this->agentLogger->info($json);
+
+        try {
+            return json_decode($json, true, 512, JSON_THROW_ON_ERROR);
+        } catch (\Throwable) {
+            return [];
+        }
+    }
+}

src/Vector/vector_ingest.py

@@ -0,0 +1,89 @@
+#!/usr/bin/env python3
+
+import sys
+import json
+from pathlib import Path
+
+# ---------------------------------------------------------
+# Argument handling
+# ---------------------------------------------------------
+if len(sys.argv) < 3:
+    print("ERROR: Missing arguments (vectorDir, knowledgeDir)")
+    sys.exit(2)
+
+vector_dir = Path(sys.argv[1]).resolve()
+knowledge_dir = Path(sys.argv[2]).resolve()
+
+index_json = knowledge_dir / "index.json"
+index_out  = vector_dir / "vector.index"
+meta_out   = vector_dir / "vector_meta.json"
+
+# ---------------------------------------------------------
+# Dependency checks
+# ---------------------------------------------------------
+try:
+    import faiss  # noqa
+except Exception:
+    print("ERROR: Python module 'faiss' not found.")
+    sys.exit(10)
+
+try:
+    from sentence_transformers import SentenceTransformer  # noqa
+except Exception:
+    print("ERROR: Python module 'sentence-transformers' not found.")
+    sys.exit(11)
+
+import faiss
+from sentence_transformers import SentenceTransformer
+
+# ---------------------------------------------------------
+# File checks
+# ---------------------------------------------------------
+if not index_json.is_file():
+    print(f"ERROR: index.json not found at {index_json}")
+    sys.exit(20)
+
+# ---------------------------------------------------------
+# Load chunks from index.json
+# ---------------------------------------------------------
+with open(index_json, "r", encoding="utf-8") as f:
+    data = json.load(f)
+
+texts = []
+ids = []
+
+for entry in data:
+    if "file" not in entry:
+        continue
+
+    chunk_path = knowledge_dir / "chunks" / entry["file"]
+    if not chunk_path.is_file():
+        continue
+
+    text = chunk_path.read_text(encoding="utf-8").strip()
+    if not text:
+        continue
+
+    texts.append(text)
+    ids.append(entry["file"])
+
+if not texts:
+    print("ERROR: No chunks loaded from index.json")
+    sys.exit(21)
+
+# ---------------------------------------------------------
+# Build vector index
+# ---------------------------------------------------------
+model = SentenceTransformer("all-MiniLM-L6-v2")
+embeddings = model.encode(texts, normalize_embeddings=True)
+
+dim = embeddings.shape[1]
+index = faiss.IndexFlatIP(dim)
+index.add(embeddings)
+
+faiss.write_index(index, str(index_out))
+
+with open(meta_out, "w", encoding="utf-8") as f:
+    json.dump(ids, f)
+
+print(f"Indexed {len(ids)} chunks.")

src/Vector/vector_search.py

@@ -0,0 +1,72 @@
+#!/usr/bin/env python3
+
+import sys
+import json
+from pathlib import Path
+
+# ---------------------------------------------------------
+# Argument handling
+# ---------------------------------------------------------
+if len(sys.argv) < 3:
+    print("ERROR: Missing arguments (query, limit)")
+    sys.exit(2)
+
+query = sys.argv[1]
+limit = int(sys.argv[2])
+
+vector_dir = Path(__file__).resolve().parent
+index_path = vector_dir / "vector.index"
+meta_path  = vector_dir / "vector_meta.json"
+
+# ---------------------------------------------------------
+# Dependency checks (controlled)
+# ---------------------------------------------------------
+try:
+    import faiss  # noqa
+except Exception:
+    print("ERROR: Python module 'faiss' not found.")
+    sys.exit(10)
+
+try:
+    from sentence_transformers import SentenceTransformer  # noqa
+except Exception:
+    print("ERROR: Python module 'sentence-transformers' not found.")
+    sys.exit(11)
+
+import faiss
+from sentence_transformers import SentenceTransformer
+
+# ---------------------------------------------------------
+# File checks
+# ---------------------------------------------------------
+if not index_path.is_file() or not meta_path.is_file():
+    print("ERROR: Vector index not found. Run vector ingest first.")
+    sys.exit(20)
+
+# ---------------------------------------------------------
+# Load model and index
+# ---------------------------------------------------------
+model = SentenceTransformer("all-MiniLM-L6-v2")
+query_vec = model.encode([query], normalize_embeddings=True)
+
+index = faiss.read_index(str(index_path))
+
+with open(meta_path, "r", encoding="utf-8") as f:
+    ids = json.load(f)
+
+# ---------------------------------------------------------
+# Search
+# ---------------------------------------------------------
+scores, indices = index.search(query_vec, limit)
+
+results = []
+for score, idx in zip(scores[0], indices[0]):
+    if idx == -1:
+        continue
+
+    results.append({
+        "chunk_id": ids[idx],
+        "score": float(score)
+    })
+
+print(json.dumps(results))