Upload app.py

app.py

@@ -48,6 +48,9 @@ sorted_by_name = []
 sorted_by_size = []
 repo_sorted_by_name = {}
 repo_sorted_by_size = {}
+initial_payload_global = None
+initial_payload_by_repo = {}
+INITIAL_PAGE_SIZE = 100
 RECORD_KEY_MAP = {
     "r": "Repo",
     "f": "File",
@@ -417,6 +420,7 @@ def load_data():
         folder_browser_data = {}
     print(f"📖 已加载 {len(records)} 条记录 ({time.time() - start:.2f}s)")
     build_indexes()
+    build_initial_payloads()
 
 def score_record(rec_idx, query_tokens, search_folders=True):
     rec = records[rec_idx]
@@ -515,6 +519,53 @@ def build_response(results: list[dict], total: int, page: int, page_size: int) -
         "page_size": page_size,
     }
 
+def trim_initial_results(raw: list[dict]) -> list[dict]:
+    keys = ("Repo", "File", "Extension", "Folder", "Size", "HasTxt")
+    return [
+        {key: record.get(key, [] if key == "Folder" else "") for key in keys}
+        for record in raw
+    ]
+
+def build_initial_payload(repo: str | None = None) -> dict:
+    if repo:
+        indices = repo_records_map.get(repo, [])
+        result_records = [records[i] for i in indices[:INITIAL_PAGE_SIZE]]
+        mode = "repo"
+        total = len(indices)
+    else:
+        result_records = records[:INITIAL_PAGE_SIZE]
+        mode = "global"
+        total = len(records)
+    return {
+        "mode": mode,
+        "repo": repo,
+        "sort": "relevance",
+        "page": 1,
+        "page_size": INITIAL_PAGE_SIZE,
+        "total": total,
+        "results": trim_initial_results(result_records),
+    }
+
+def build_initial_payloads():
+    global initial_payload_global, initial_payload_by_repo
+    initial_payload_global = build_initial_payload(None)
+    initial_payload_by_repo = {
+        repo: build_initial_payload(repo)
+        for repo in repo_records_map
+    }
+
+def inject_initial_payload(html: str, repo_short: str | None = None) -> str:
+    repo = f"VoiceOfML/{repo_short}" if repo_short else None
+    payload = initial_payload_by_repo.get(repo) if repo else initial_payload_global
+    if not payload:
+        return html
+    script = (
+        '<script id="initial-search-data" type="application/json">'
+        + json.dumps(payload, ensure_ascii=False, separators=(",", ":")).replace("</", "<\\/")
+        + "</script>"
+    )
+    return html.replace("</body>", script + "\n</body>")
+
 def apply_mixed_folder_filters(indices, self_folders=None, subtree_folders=None):
     self_set = {str(path).strip("/") for path in (self_folders or []) if str(path).strip("/")}
     subtree_set = {str(path).strip("/") for path in (subtree_folders or []) if str(path).strip("/")}
@@ -956,10 +1007,21 @@ def serve_sw():
     if p.exists():
         return PlainTextResponse(p.read_text(encoding="utf-8"), media_type="application/javascript")
     return PlainTextResponse("", status_code=404)
+
+@app.get("/")
+async def serve_root():
+    index_path = Path("static/index.html")
+    if index_path.exists():
+        return HTMLResponse(inject_initial_payload(index_path.read_text(encoding="utf-8"), None))
+    return HTMLResponse("<h1>VoiceOfML Search</h1>", status_code=200)
+
 @app.get("/{rest_of_path:path}")
 
 async def serve_spa(rest_of_path: str):
     index_path = Path("static/index.html")
     if index_path.exists():
-        return HTMLResponse(index_path.read_text(encoding="utf-8"))
+        repo_short = rest_of_path.strip("/").split("/")[0] if rest_of_path.strip("/") else None
+        if repo_short and f"VoiceOfML/{repo_short}" not in repo_counts:
+            repo_short = None
+        return HTMLResponse(inject_initial_payload(index_path.read_text(encoding="utf-8"), repo_short))
     return HTMLResponse("<h1>VoiceOfML Search</h1>", status_code=200)