Processing...
-
diff --git a/frontend/vite.config.js b/frontend/vite.config.js
index a52ca07..1855f65 100644
--- a/frontend/vite.config.js
+++ b/frontend/vite.config.js
@@ -1,48 +1,48 @@
-import { defineConfig } from 'vite'
-import react from '@vitejs/plugin-react'
-import path from 'path'
-import { fileURLToPath } from 'url'
-
-const __dirname = path.dirname(fileURLToPath(import.meta.url))
+import { defineConfig } from "vite";
+import react from "@vitejs/plugin-react";
+import path from "path";
+import { fileURLToPath } from "url";
+const __dirname = path.dirname(fileURLToPath(import.meta.url));
+const baseUrl = "http://localhost:8000";
+console.log("Using backend URL:", baseUrl);
export default defineConfig({
plugins: [react()],
resolve: {
alias: {
- '@': path.resolve(__dirname, './src'),
+ "@": path.resolve(__dirname, "./src"),
},
},
build: {
- outDir: '../static',
+ outDir: "../static",
emptyOutDir: true,
},
server: {
proxy: {
- '/libraries': {
- target: 'http://localhost:8000',
+ "/libraries": {
+ target: baseUrl,
changeOrigin: true,
},
- '/convert': {
- target: 'http://localhost:8000',
+ "/convert": {
+ target: baseUrl,
changeOrigin: true,
},
- '/compare': {
- target: 'http://localhost:8000',
+ "/compare": {
+ target: baseUrl,
changeOrigin: true,
},
- '/history': {
- target: 'http://localhost:8000',
+ "/history": {
+ target: baseUrl,
changeOrigin: true,
},
- '/stats': {
- target: 'http://localhost:8000',
+ "/stats": {
+ target: baseUrl,
changeOrigin: true,
},
- '/health': {
- target: 'http://localhost:8000',
+ "/health": {
+ target: baseUrl,
changeOrigin: true,
},
},
},
-})
-
+});
diff --git a/main.py b/main.py
index 43bbf0a..f1aec86 100644
--- a/main.py
+++ b/main.py
@@ -38,11 +38,10 @@
UPLOAD_DIR = "uploads"
os.makedirs(UPLOAD_DIR, exist_ok=True)
-# Initialize factory (singleton pattern)
-factory = OCRFactory()
-
# Initialize comparison services
db_service = DatabaseService()
+# Initialize factory (singleton pattern) with DB service
+factory = OCRFactory(db_service)
queue_manager = QueueManager(db_service)
results_manager = ResultsManager()
@@ -68,6 +67,30 @@ async def get_available_libraries():
"""Get list of available conversion libraries"""
return {"libraries": factory.list_all_converters()}
+@app.post("/libraries/status")
+async def set_library_status(
+ library_name: str = Form(...),
+ is_enabled: bool = Form(...)
+):
+ """Enable or disable a library"""
+ try:
+ # Update DB
+ db_service.set_library_status(library_name, is_enabled)
+
+ # Refresh factory
+ factory.refresh_converters()
+
+ return {
+ "status": "success",
+ "library": library_name,
+ "enabled": is_enabled,
+ "message": f"Library {library_name} {'enabled' if is_enabled else 'disabled'}"
+ }
+ except Exception as e:
+ logger.error(f"Error setting library status: {e}")
+ raise HTTPException(status_code=500, detail=str(e))
+
+
@app.post("/convert")
async def convert_pdf(
file: UploadFile = File(...),
diff --git a/services/db_methods_snippet.py b/services/db_methods_snippet.py
new file mode 100644
index 0000000..bf61a04
--- /dev/null
+++ b/services/db_methods_snippet.py
@@ -0,0 +1,25 @@
+
+ def get_enabled_libraries(self):
+ """Get list of enabled library names"""
+ conn = sqlite3.connect(self.db_path, timeout=30.0)
+ cursor = conn.cursor()
+
+ cursor.execute('SELECT library_name FROM user_libraries WHERE is_enabled = 1')
+ libraries = [row[0] for row in cursor.fetchall()]
+
+ conn.close()
+ return libraries
+
+ def set_library_status(self, library_name, is_enabled):
+ """Enable or disable a library"""
+ conn = sqlite3.connect(self.db_path, timeout=30.0)
+ cursor = conn.cursor()
+
+ cursor.execute('''
+ INSERT INTO user_libraries (library_name, is_enabled)
+ VALUES (?, ?)
+ ON CONFLICT(library_name) DO UPDATE SET is_enabled = ?
+ ''', (library_name, 1 if is_enabled else 0, 1 if is_enabled else 0))
+
+ conn.commit()
+ conn.close()
diff --git a/services/db_service.py b/services/db_service.py
index 899f690..d348d00 100644
--- a/services/db_service.py
+++ b/services/db_service.py
@@ -49,10 +49,36 @@ def init_db(self):
CONSTRAINT status_check CHECK (status IN ('pending', 'success', 'failed', 'timeout'))
)
''')
+
+ # User-opted libraries table
+ cursor.execute('''
+ CREATE TABLE IF NOT EXISTS user_libraries (
+ library_name TEXT PRIMARY KEY,
+ is_enabled BOOLEAN DEFAULT 0
+ )
+ ''')
# Create indexes for faster queries
cursor.execute('CREATE INDEX IF NOT EXISTS idx_tasks_created ON tasks(created_at DESC)')
cursor.execute('CREATE INDEX IF NOT EXISTS idx_comparisons_task ON comparisons(task_id)')
+
+ # Seed initial libraries if empty
+ cursor.execute('SELECT COUNT(*) FROM user_libraries')
+ if cursor.fetchone()[0] == 0:
+ initial_libraries = [
+ ('pymupdf4llm', 0),
+ ('markitdown', 0),
+ ('marker', 0),
+ ('docling', 0),
+ ('paddleocr', 0),
+ ('deepseekocr', 0),
+ ('pytesseract', 0),
+ ('unstructured', 0),
+ ]
+ cursor.executemany(
+ 'INSERT INTO user_libraries (library_name, is_enabled) VALUES (?, ?)',
+ initial_libraries
+ )
conn.commit()
conn.close()
@@ -201,6 +227,7 @@ def get_library_stats(self):
conn.close()
return stats
+
def delete_task(self, task_id):
"""Delete task and its comparisons"""
conn = sqlite3.connect(self.db_path, timeout=30.0)
@@ -214,3 +241,29 @@ def delete_task(self, task_id):
conn.commit()
conn.close()
+ def get_enabled_libraries(self):
+ """Get list of enabled library names"""
+ conn = sqlite3.connect(self.db_path, timeout=30.0)
+ cursor = conn.cursor()
+
+ cursor.execute('SELECT library_name FROM user_libraries WHERE is_enabled = 1')
+ libraries = [row[0] for row in cursor.fetchall()]
+
+ conn.close()
+ return libraries
+
+ def set_library_status(self, library_name, is_enabled):
+ """Enable or disable a library"""
+ conn = sqlite3.connect(self.db_path, timeout=30.0)
+ cursor = conn.cursor()
+
+ cursor.execute('''
+ INSERT INTO user_libraries (library_name, is_enabled)
+ VALUES (?, ?)
+ ON CONFLICT(library_name) DO UPDATE SET is_enabled = ?
+ ''', (library_name, 1 if is_enabled else 0, 1 if is_enabled else 0))
+
+ conn.commit()
+ conn.close()
+
+
diff --git a/services/ocrfactory.py b/services/ocrfactory.py
index 878eb47..2a2afab 100644
--- a/services/ocrfactory.py
+++ b/services/ocrfactory.py
@@ -16,13 +16,37 @@ class OCRFactory:
Implements singleton pattern for converter instances.
"""
- def __init__(self):
+ def __init__(self, db_service=None):
self._converters: Dict[str, PDFConverter] = {}
+ # If no DB service provided, create one
+ if db_service is None:
+ from services.db_service import DatabaseService
+ self.db_service = DatabaseService()
+ else:
+ self.db_service = db_service
+
+ self._register_default_converters()
+
+ def refresh_converters(self):
+ """Reload converters from database settings"""
+ self._converters = {}
self._register_default_converters()
def _register_default_converters(self):
- """Register all available converter implementations"""
- converters = [
+ """Register all available converter implementations based on DB settings"""
+ # Get enabled libraries from DB
+ try:
+ enabled_libraries = set(self.db_service.get_enabled_libraries())
+ except Exception as e:
+ logger.error(f"Failed to load enabled libraries from DB: {e}")
+
+ logger.info(f"Loading enabled libraries: {enabled_libraries}")
+
+ if len(enabled_libraries) == 0:
+ logger.warning("No enabled libraries found in DB. Registering all converters.")
+ return
+
+ all_converters = [
PyMuPDF4LLMConverter(),
MarkItDownConverter(),
MarkerConverter(),
@@ -33,7 +57,11 @@ def _register_default_converters(self):
UnstructuredConverter(),
]
- for converter in converters:
+ for converter in all_converters:
+ # Skip if not enabled in DB
+ if converter.name not in enabled_libraries:
+ continue
+
if converter.available:
self._converters[converter.name] = converter
logger.info(f"Registered converter: {converter.name}")
@@ -52,6 +80,11 @@ def list_available_converters(self) -> List[str]:
def list_all_converters(self) -> List[Dict[str, bool]]:
"""List all converters with their availability status"""
+
+ try:
+ enabled_libraries = set(self.db_service.get_enabled_libraries())
+ except Exception as e:
+ logger.error(f"Failed to load enabled libraries from DB: {e}")
all_converters = [
"pymupdf4llm",
"markitdown",
@@ -70,7 +103,8 @@ def list_all_converters(self) -> List[Dict[str, bool]]:
result.append({
"name": name,
"available": available,
- "error": None if available else getattr(converter, "error_message", "Unavailable")
+ "error": None if available else getattr(converter, "error_message", "Unavailable"),
+ "enabled": name in enabled_libraries
})
return result