Upload 2 files

.gitattributes

@@ -37,3 +37,4 @@ mub_chroma_db/chroma.sqlite3 filter=lfs diff=lfs merge=lfs -text
 chroma_db/chroma.sqlite3 filter=lfs diff=lfs merge=lfs -text
 mub_chroma_db_distilbert/chroma.sqlite3 filter=lfs diff=lfs merge=lfs -text
 mub_chroma_db/** -filter -diff -merge -text
+mub_chroma_db.bundle filter=lfs diff=lfs merge=lfs -text

app.py

@@ -2,7 +2,10 @@ import logging
 import math
 import os
 import re
+import shutil
+import zipfile
 from collections import Counter, defaultdict
+from pathlib import Path
 from typing import Dict, List
 
 os.environ.setdefault("ANONYMIZED_TELEMETRY", "False")
@@ -29,6 +32,9 @@ VECTOR_CANDIDATES = 10
 LEXICAL_CANDIDATES = 10
 CONTEXT_RESULTS = 5
 RRF_K = 60
+DATABASE_BUNDLE = Path(__file__).with_name("mub_chroma_db.bundle")
+DATABASE_RUNTIME_PATH = Path("/tmp/mub_chroma_db")
+EXPECTED_DATABASE_RECORDS = 223
 
 
 def normalize(text: str) -> str:
@@ -61,6 +67,20 @@ def detect_language(text: str) -> str:
     return "ar" if total_chars and arabic_chars / total_chars > 0.3 else "en"
 
 
+def materialize_database() -> str:
+    sqlite_path = DATABASE_RUNTIME_PATH / "mub_chroma_db" / "chroma.sqlite3"
+    if not sqlite_path.exists():
+        if not DATABASE_BUNDLE.exists():
+            raise RuntimeError(f"Missing database bundle: {DATABASE_BUNDLE}")
+        shutil.rmtree(DATABASE_RUNTIME_PATH, ignore_errors=True)
+        DATABASE_RUNTIME_PATH.mkdir(parents=True)
+        with zipfile.ZipFile(DATABASE_BUNDLE) as archive:
+            archive.extractall(DATABASE_RUNTIME_PATH)
+    if sqlite_path.read_bytes()[:16] != b"SQLite format 3\x00":
+        raise RuntimeError("Bundled Chroma database is not a valid SQLite file.")
+    return str(sqlite_path.parent)
+
+
 class BM25Index:
     def __init__(self, documents: List[str], k1: float = 1.5, b: float = 0.75):
         self.k1 = k1
@@ -145,12 +165,18 @@ class FieldedBM25:
 class SimpleRAG:
     """Targeted Arabert retrieval fused with field-aware BM25."""
 
-    def __init__(self, db_path: str = "./mub_chroma_db"):
+    def __init__(self, db_path: str = None):
+        db_path = db_path or materialize_database()
         logger.info("Loading Arabert embedding model...")
         self.model = SentenceTransformer(EMBEDDING_MODEL)
         logger.info("Loading targeted ChromaDB from %s...", db_path)
         self.client = chromadb.PersistentClient(path=db_path)
         self.collection = self.client.get_collection(name="mub_info")
+        if self.collection.count() != EXPECTED_DATABASE_RECORDS:
+            raise RuntimeError(
+                "Unexpected database record count: "
+                f"{self.collection.count()} != {EXPECTED_DATABASE_RECORDS}"
+            )
         corpus = self.collection.get(include=["documents", "metadatas"])
         self.records = []
         for chroma_id, document, metadata in zip(

mub_chroma_db.bundle

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fbb68e81b701dcfcd3fa2839cd4b0868412db90450db86891eb7c190cf07640
+size 1422842