Upload app.py

app.py

@@ -104,9 +104,12 @@ def build_indexes() -> None:
     for idx, rec in enumerate(records):
         source_records_map.setdefault(rec["source"], []).append(idx)
         extension_counts[rec["extension"]] = extension_counts.get(rec["extension"], 0) + 1
-        search_text = " ".join([rec["display_name"], rec["display_rel_path"], rec["source_name"]])
-        tokens = tokenize(search_text)
-        rec["_search_text"] = search_text.lower()
+        file_text = " ".join([rec["display_name"], rec["source_name"]])
+        path_text = rec["display_rel_path"]
+        tokens = tokenize(f"{file_text} {path_text}")
+        rec["_file_search_text"] = file_text.lower()
+        rec["_path_search_text"] = path_text.lower()
+        rec["_search_text"] = f"{file_text} {path_text}".lower()
         for token in tokens:
             word_index.setdefault(token, set()).add(idx)
             did_you_mean_vocab[token] = did_you_mean_vocab.get(token, 0) + 1
@@ -128,12 +131,15 @@ def load_data() -> None:
     print(f"loaded {len(records)} txt records in {time.time() - start:.2f}s")
 
 
-def score_record(idx: int, query_tokens: list[str]) -> int:
+def score_record(idx: int, query_tokens: list[str], search_paths: bool = True) -> int:
     score = 0
-    text = records[idx]["_search_text"]
+    file_text = records[idx]["_file_search_text"]
+    path_text = records[idx]["_path_search_text"]
     for token in query_tokens:
-        if token in text:
+        if token in file_text:
             score += 3
+        if search_paths and token in path_text:
+            score += 2
     return score
 
 
@@ -171,7 +177,7 @@ def trim_record(rec: dict) -> dict:
     }
 
 
-def search(q="", sources_filter=None, folders=None, min_size=None, max_size=None, page=1, page_size=100, sort="relevance", exact=False):
+def search(q="", sources_filter=None, folders=None, min_size=None, max_size=None, page=1, page_size=100, sort="relevance", exact=False, search_paths=True):
     q = q.strip()
     if not q:
         indices = list(range(len(records)))
@@ -179,24 +185,33 @@ def search(q="", sources_filter=None, folders=None, min_size=None, max_size=None
     else:
         tokens = tokenize(q)
         if exact:
-            indices = [idx for idx, rec in enumerate(records) if q.lower() in rec["_search_text"]]
+            if search_paths:
+                indices = [idx for idx, rec in enumerate(records) if q.lower() in rec["_search_text"]]
+            else:
+                indices = [idx for idx, rec in enumerate(records) if q.lower() in rec["_file_search_text"]]
         else:
             matched = set()
             for token in tokens:
                 if token in word_index:
-                    matched.update(word_index[token])
+                    if search_paths:
+                        matched.update(word_index[token])
+                    else:
+                        matched.update(idx for idx in word_index[token] if token in records[idx]["_file_search_text"])
                     continue
                 tok_len = len(token)
                 for delta in (-2, -1, 0, 1, 2):
                     for vocab_word in vocab_by_len.get(tok_len + delta, {}):
                         if edit_distance(token, vocab_word) <= 2:
-                            matched.update(word_index.get(vocab_word, set()))
+                            if search_paths:
+                                matched.update(word_index.get(vocab_word, set()))
+                            else:
+                                matched.update(idx for idx in word_index.get(vocab_word, set()) if vocab_word in records[idx]["_file_search_text"])
                             if len(matched) > 5000:
                                 break
             indices = list(matched)
         filtered = apply_filters(indices, sources_filter, folders, min_size, max_size)
         if sort == "relevance":
-            filtered.sort(key=lambda idx: (-score_record(idx, tokens), records[idx]["display_rel_path"].lower()))
+            filtered.sort(key=lambda idx: (-score_record(idx, tokens, search_paths), records[idx]["display_rel_path"].lower()))
 
     if sort == "name":
         filtered.sort(key=lambda idx: records[idx]["display_rel_path"].lower())
@@ -243,6 +258,7 @@ class SearchRequest(BaseModel):
     page_size: int = 100
     sort: str = "relevance"
     exact: bool = False
+    search_paths: bool = True
 
 
 class ZipRequest(BaseModel):
@@ -251,14 +267,14 @@ class ZipRequest(BaseModel):
 
 @app.post("/api/search")
 def api_search(body: SearchRequest):
-    return JSONResponse(search(body.q, body.sources, body.folders, body.min_size, body.max_size, body.page, body.page_size, body.sort, body.exact))
+    return JSONResponse(search(body.q, body.sources, body.folders, body.min_size, body.max_size, body.page, body.page_size, body.sort, body.exact, body.search_paths))
 
 
 @app.post("/api/search/{source_slug}")
 def api_search_source(source_slug: str, body: SearchRequest):
     if source_slug not in source_counts:
         return JSONResponse({"error": "source not found", "results": [], "total": 0}, status_code=404)
-    return JSONResponse(search(body.q, [source_slug], body.folders, body.min_size, body.max_size, body.page, body.page_size, body.sort, body.exact))
+    return JSONResponse(search(body.q, [source_slug], body.folders, body.min_size, body.max_size, body.page, body.page_size, body.sort, body.exact, body.search_paths))
 
 
 @app.get("/api/sources")