v0.0.2: split api_bp into upload/process/download modules, Connection: close, upload tests (13 new, 106 total)

site/app.py

@@ -20,7 +20,7 @@ if _sys_path_root not in sys.path:
     sys.path.insert(0, _sys_path_root)
 
 # Версия приложения (меняется при изменениях)
-VERSION = "0.0.1"
+VERSION = "0.0.2"
 
 
 def create_app():
@@ -51,6 +51,7 @@ def create_app():
         response.headers["Cache-Control"] = "no-cache, no-store, must-revalidate"
         response.headers["Pragma"] = "no-cache"
         response.headers["Expires"] = "0"
+        response.headers["Connection"] = "close"
         return response
 
     return app

site/routes/__init__.py

@@ -1,20 +1,13 @@
-"""
-Регистрация всех blueprint'ов приложения.
-
-Импортируется из site/app.py при создании Flask-приложения.
-"""
-
 from .main_bp import main_bp
 from .health_bp import health_bp
-from .api_bp import api_bp
+from .upload_bp import upload_bp
+from .process_bp import process_bp
+from .download_bp import download_bp
 
 
 def register_routes(app):
-    """Зарегистрировать все blueprint'ы на Flask-приложении.
-
-    Args:
-        app: Экземпляр Flask-приложения
-    """
     app.register_blueprint(main_bp)
     app.register_blueprint(health_bp)
-    app.register_blueprint(api_bp)
+    app.register_blueprint(upload_bp)
+    app.register_blueprint(process_bp)
+    app.register_blueprint(download_bp)

site/routes/api_bp.py

@@ -1,83 +0,0 @@
-"""
-Blueprint: API DrHider.
-
-Three endpoints:
-- POST /api/upload      — upload one file -> {session_id}
-- POST /api/process/{sid} — process all session files -> {status:"done"}
-- GET  /api/download/{sid} — download ZIP
-"""
-
-import io
-import zipfile
-import traceback
-from flask import Blueprint, request, send_file, jsonify
-
-from drhider import obfuscate_files, LLMClient
-from drhider.builder import build_zip, build_mapping_csv
-from session import create_session, add_file, get_files, store_result, get_result, cleanup, file_count
-
-api_bp = Blueprint("api", __name__, url_prefix="/api")
-
-
-@api_bp.route("/upload", methods=["POST"])
-def upload():
-    """Upload one file to session."""
-    sid = request.form.get("session", "")
-    if not sid:
-        sid = create_session()
-    uploaded = request.files.getlist("files")
-    if not uploaded:
-        return jsonify({"ok": False, "error": "No file"}), 400
-    f = uploaded[0]
-    if not f.filename:
-        return jsonify({"ok": False, "error": "No filename"}), 400
-    ok = add_file(sid, f.filename, f.read())
-    if not ok:
-        return jsonify({"ok": False, "error": "Session not found"}), 404
-    return jsonify({"ok": True, "session": sid, "count": file_count(sid)})
-
-
-@api_bp.route("/process/<sid>", methods=["POST"])
-def process(sid):
-    """Process all session files one by one -> ZIP."""
-    files = get_files(sid)
-    if files is None:
-        return jsonify({"ok": False, "error": "Session not found"}), 404
-    if not files:
-        return jsonify({"ok": False, "error": "No files"}), 400
-    try:
-        llm = LLMClient()
-        all_results = []
-        all_mapping = {}
-        for fname, content in files:
-            zip_data, _csv = obfuscate_files([(fname, content, "")], llm_client=llm)
-            with zipfile.ZipFile(io.BytesIO(zip_data)) as zf:
-                for name in zf.namelist():
-                    if name == "mapping.csv":
-                        csv_text = zf.read(name).decode("utf-8")
-                        lines = csv_text.strip().split("\n")
-                        if len(lines) > 1:
-                            for line in lines[1:]:
-                                parts = line.split(",", 2)
-                                if len(parts) >= 3:
-                                    all_mapping[parts[1]] = parts[2]
-                    elif not name.endswith("/"):
-                        all_results.append((name, zf.read(name)))
-        csv_str = build_mapping_csv(all_mapping) if all_mapping else ""
-        final_zip = build_zip(all_results, mapping_csv=csv_str)
-        store_result(sid, final_zip)
-        return jsonify({"ok": True, "status": "done", "files": len(all_results)})
-    except Exception as e:
-        traceback.print_exc()
-        return jsonify({"ok": False, "error": str(e)}), 500
-
-
-@api_bp.route("/download/<sid>", methods=["GET"])
-def download(sid):
-    """Download result and cleanup session."""
-    zip_data = get_result(sid)
-    if zip_data is None:
-        return jsonify({"ok": False, "error": "Not found"}), 404
-    cleanup(sid)
-    return send_file(io.BytesIO(zip_data), mimetype="application/zip",
-                     as_attachment=True, download_name="drhider_output.zip")

site/routes/download_bp.py

@@ -0,0 +1,21 @@
+"""
+GET /api/download/{sid} — скачать ZIP и удалить сессию.
+
+После этого вызова — всё чисто, памяти ноль.
+"""
+
+import io
+from flask import Blueprint, send_file, jsonify
+from session import get_result, cleanup
+
+download_bp = Blueprint("download", __name__, url_prefix="/api")
+
+
+@download_bp.route("/download/<sid>", methods=["GET"])
+def download(sid):
+    zip_data = get_result(sid)
+    if zip_data is None:
+        return jsonify({"ok": False, "error": "Not found"}), 404
+    cleanup(sid)
+    return send_file(io.BytesIO(zip_data), mimetype="application/zip",
+                     as_attachment=True, download_name="drhider_output.zip")

site/routes/process_bp.py

@@ -0,0 +1,47 @@
+"""
+POST /api/process/{sid} — обработать ВСЕ файлы сессии по одному.
+
+Блокирует до завершения. Результат сохраняется в сессии.
+"""
+
+import io, zipfile, traceback
+from flask import Blueprint, jsonify
+from drhider import obfuscate_files, LLMClient
+from drhider.builder import build_zip, build_mapping_csv
+from session import get_files, store_result
+
+process_bp = Blueprint("process", __name__, url_prefix="/api")
+
+
+@process_bp.route("/process/<sid>", methods=["POST"])
+def process(sid):
+    files = get_files(sid)
+    if files is None:
+        return jsonify({"ok": False, "error": "Session not found"}), 404
+    if not files:
+        return jsonify({"ok": False, "error": "No files"}), 400
+    try:
+        llm = LLMClient()
+        all_results = []
+        all_mapping = {}
+        for fname, content in files:
+            zip_data, _csv = obfuscate_files([(fname, content, "")], llm_client=llm)
+            with zipfile.ZipFile(io.BytesIO(zip_data)) as zf:
+                for name in zf.namelist():
+                    if name == "mapping.csv":
+                        csv_text = zf.read(name).decode("utf-8")
+                        lines = csv_text.strip().split("\n")
+                        if len(lines) > 1:
+                            for line in lines[1:]:
+                                parts = line.split(",", 2)
+                                if len(parts) >= 3:
+                                    all_mapping[parts[1]] = parts[2]
+                    elif not name.endswith("/"):
+                        all_results.append((name, zf.read(name)))
+        csv_str = build_mapping_csv(all_mapping) if all_mapping else ""
+        final_zip = build_zip(all_results, mapping_csv=csv_str)
+        store_result(sid, final_zip)
+        return jsonify({"ok": True, "status": "done", "files": len(all_results)})
+    except Exception as e:
+        traceback.print_exc()
+        return jsonify({"ok": False, "error": str(e)}), 500

site/routes/upload_bp.py

@@ -0,0 +1,28 @@
+"""
+POST /api/upload — загрузить ОДИН файл.
+
+Каждый вызов — новое соединение, передал → закрыл.
+Сессия связывает файлы через session_id.
+"""
+
+from flask import Blueprint, request, jsonify
+from session import create_session, add_file, file_count
+
+upload_bp = Blueprint("upload", __name__, url_prefix="/api")
+
+
+@upload_bp.route("/upload", methods=["POST"])
+def upload():
+    sid = request.form.get("session", "")
+    if not sid:
+        sid = create_session()
+    uploaded = request.files.getlist("files")
+    if not uploaded:
+        return jsonify({"ok": False, "error": "No file"}), 400
+    f = uploaded[0]
+    if not f.filename:
+        return jsonify({"ok": False, "error": "No filename"}), 400
+    ok = add_file(sid, f.filename, f.read())
+    if not ok:
+        return jsonify({"ok": False, "error": "Session not found"}), 404
+    return jsonify({"ok": True, "session": sid, "count": file_count(sid)})

tests/test_upload.py

@@ -0,0 +1,283 @@
+"""
+Тесты POST /api/upload — загрузка файлов в сессию.
+
+Запуск:
+    python3 tests/test_upload.py
+"""
+
+import io
+import os
+import sys
+import time
+import json
+import gc
+
+# Настраиваем пути
+_site = os.path.join(os.path.dirname(__file__), "..", "site")
+sys.path.insert(0, _site)
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+
+from flask import Flask
+
+# Импортируем напрямую, минуя routes/__init__.py (process_bp тянет drhider)
+import importlib.util
+_spec = importlib.util.spec_from_file_location(
+    "upload_bp", os.path.join(_site, "routes", "upload_bp.py")
+)
+_upload_mod = importlib.util.module_from_spec(_spec)
+_spec.loader.exec_module(_upload_mod)
+upload_bp = _upload_mod.upload_bp
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Helpers
+# ═══════════════════════════════════════════════════════════════════════════
+
+def _new_client():
+    """Создать свежий тестовый клиент."""
+    app = Flask(__name__)
+    app.register_blueprint(upload_bp)
+    return app.test_client()
+
+
+def _get_rss():
+    """RSS процесса в байтах (Linux)."""
+    try:
+        with open("/proc/self/statm", "r") as f:
+            parts = f.read().split()
+            return int(parts[1]) * 4096
+    except Exception:
+        return 0
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Базовые тесты
+# ═══════════════════════════════════════════════════════════════════════════
+
+def test_no_file():
+    """POST без файла → 400."""
+    c = _new_client()
+    rv = c.post("/api/upload")
+    assert rv.status_code == 400
+    data = json.loads(rv.data)
+    assert data["ok"] is False
+    assert "No file" in data["error"]
+
+
+def test_no_filename():
+    """POST с файлом без имени → 400."""
+    c = _new_client()
+    data = {"files": (io.BytesIO(b"hello"), "")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 400
+    resp = json.loads(rv.data)
+    assert resp["ok"] is False
+    assert "filename" in resp["error"].lower()
+
+
+def test_single_file():
+    """POST с одним файлом → 200, сессия + count=1."""
+    c = _new_client()
+    data = {"files": (io.BytesIO(b"hello world"), "test.txt")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 200
+    resp = json.loads(rv.data)
+    assert resp["ok"] is True
+    assert len(resp["session"]) == 12
+    assert resp["count"] == 1
+
+
+def test_two_files_same_session():
+    """Два файла в одну сессию → count=2."""
+    c = _new_client()
+    data1 = {"files": (io.BytesIO(b"first"), "file1.txt")}
+    rv1 = c.post("/api/upload", data=data1, content_type="multipart/form-data")
+    assert rv1.status_code == 200
+    sid = json.loads(rv1.data)["session"]
+
+    data2 = {
+        "files": (io.BytesIO(b"second"), "file2.txt"),
+        "session": sid,
+    }
+    rv2 = c.post("/api/upload", data=data2, content_type="multipart/form-data")
+    assert rv2.status_code == 200
+    resp2 = json.loads(rv2.data)
+    assert resp2["session"] == sid
+    assert resp2["count"] == 2
+
+
+def test_fake_session_404():
+    """Несуществующая сессия → 404."""
+    c = _new_client()
+    data = {
+        "files": (io.BytesIO(b"data"), "f.txt"),
+        "session": "deadbeef1234",
+    }
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 404
+    resp = json.loads(rv.data)
+    assert resp["ok"] is False
+
+
+def test_empty_file():
+    """Пустой файл (0 байт) → 200."""
+    c = _new_client()
+    data = {"files": (io.BytesIO(b""), "empty.txt")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 200
+    resp = json.loads(rv.data)
+    assert resp["ok"] is True
+    assert resp["count"] == 1
+
+
+def test_cyrillic_filename():
+    """Кириллическое имя файла → 200."""
+    c = _new_client()
+    content = "кириллица".encode("utf-8")
+    data = {"files": (io.BytesIO(content), "договор_№123.txt")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 200
+    resp = json.loads(rv.data)
+    assert resp["ok"] is True
+    assert resp["count"] == 1
+
+
+def test_content_type_json():
+    """Ответ — application/json."""
+    c = _new_client()
+    data = {"files": (io.BytesIO(b"x"), "f.txt")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 200
+    assert "application/json" in rv.content_type
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Большие файлы
+# ═══════════════════════════════════════════════════════════════════════════
+
+def test_large_10mb():
+    """Файл 10 MB — загружается без ошибок."""
+    c = _new_client()
+    size = 10 * 1024 * 1024
+    big_data = b"A" * size
+    data = {"files": (io.BytesIO(big_data), "big_10mb.bin")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 200
+    resp = json.loads(rv.data)
+    assert resp["ok"] is True
+    assert resp["count"] == 1
+
+
+def test_large_50mb():
+    """Файл 50 MB — загружается без ошибок."""
+    c = _new_client()
+    size = 50 * 1024 * 1024
+    big_data = b"B" * size
+    data = {"files": (io.BytesIO(big_data), "big_50mb.bin")}
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    assert rv.status_code == 200
+    resp = json.loads(rv.data)
+    assert resp["ok"] is True
+    assert resp["count"] == 1
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Много мелких файлов
+# ═══════════════════════════════════════════════════════════════════════════
+
+def test_many_small_files():
+    """100 файлов по 1 KB в одну сессию."""
+    c = _new_client()
+    data1 = {"files": (io.BytesIO(b"x" * 1024), "file_000.txt")}
+    rv1 = c.post("/api/upload", data=data1, content_type="multipart/form-data")
+    sid = json.loads(rv1.data)["session"]
+
+    for i in range(1, 100):
+        data = {
+            "files": (io.BytesIO(b"y" * 1024), f"file_{i:03d}.txt"),
+            "session": sid,
+        }
+        rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+        assert rv.status_code == 200
+        resp = json.loads(rv.data)
+        assert resp["session"] == sid
+        assert resp["count"] == i + 1
+
+    assert resp["count"] == 100
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Утечка памяти
+# ═══════════════════════════════════════════════════════════════════════════
+
+def test_memory_after_large_files():
+    """После загрузки больших файлов память не течёт."""
+    c = _new_client()
+    gc.collect()
+    mem_before = _get_rss()
+
+    for i in range(10):
+        data = {"files": (io.BytesIO(b"M" * (1024 * 1024)), f"memtest_{i}.bin")}
+        rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+        assert rv.status_code == 200
+
+    gc.collect()
+    mem_after = _get_rss()
+    diff_mb = (mem_after - mem_before) / (1024 * 1024)
+    assert diff_mb < 50, f"Memory grew {diff_mb:.1f} MB, expected < 50 MB"
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Бенчмарк скорости
+# ═══════════════════════════════════════════════════════════════════════════
+
+def test_upload_speed():
+    """1 MB файл — загрузка < 500 ms."""
+    c = _new_client()
+    data = {"files": (io.BytesIO(b"S" * (1024 * 1024)), "speed.bin")}
+
+    start = time.time()
+    rv = c.post("/api/upload", data=data, content_type="multipart/form-data")
+    elapsed = time.time() - start
+
+    assert rv.status_code == 200
+    assert elapsed < 0.5, f"Upload took {elapsed:.3f}s, expected < 0.5s"
+
+
+# ═══════════════════════════════════════════════════════════════════════════
+# Основной блок
+# ═══════════════════════════════════════════════════════════════════════════
+
+if __name__ == "__main__":
+    passed = 0
+    failed = 0
+    tests = [
+        test_no_file,
+        test_no_filename,
+        test_single_file,
+        test_two_files_same_session,
+        test_fake_session_404,
+        test_empty_file,
+        test_cyrillic_filename,
+        test_content_type_json,
+        test_large_10mb,
+        test_large_50mb,
+        test_many_small_files,
+        test_memory_after_large_files,
+        test_upload_speed,
+    ]
+    for test in tests:
+        try:
+            test()
+            print(f"  ✅ {test.__name__}")
+            passed += 1
+        except Exception as e:
+            print(f"  ❌ {test.__name__}: {e}")
+            failed += 1
+    print(f"\n{'='*50}")
+    print(f"Результат: {passed} OK, {failed} FAIL из {len(tests)}")
+    if failed == 0:
+        print("✅ ВСЕ ТЕСТЫ ПРОЙДЕНЫ")
+    else:
+        print("❌ ЕСТЬ ПРОВАЛЫ")
+        sys.exit(1)