fix bugs

dedoc/config.py

@@ -24,7 +24,7 @@
 
     # --------------------------------------------GPU SETTINGS----------------------------------------------------------
     # set gpu in XGBoost and torch models
-    on_gpu=True,
+    on_gpu=False,
 
     # ---------------------------------------------API SETTINGS---------------------------------------------------------
     # max file size in bytes

docker-compose.yml

@@ -12,8 +12,7 @@ services:
       - 1231:1231
     environment:
       DOCREADER_PORT: 1231
-      NVIDIA_VISIBLE_DEVICES: 0
-    runtime: nvidia
+
 
   test:
     depends_on:
@@ -27,8 +26,6 @@ services:
       DOCREADER_PORT: 1231
       is_test: $test
       PYTHONPATH: $PYTHONPATH:/dedoc_root/tests:/dedoc_root
-      NVIDIA_VISIBLE_DEVICES: 0
-    runtime: nvidia
     command:
         bash dedoc_root/tests/run_tests_in_docker.sh
 

docker_gpu/Dockerfile

@@ -4,14 +4,15 @@ FROM dedocproject/dedoc_p3.9_base:version_2023_08_28
 ENV PYTHONPATH "${PYTHONPATH}:/dedoc_root"
 ENV RESOURCES_PATH "/dedoc_root/resources"
 
+
 ADD requirements.txt .
 RUN pip3 install --no-cache-dir -r requirements.txt
 RUN pip install torch==1.11.0+cu113 torchvision==0.12.0+cu113 -f https://download.pytorch.org/whl/torch_stable.html
 
 
 RUN mkdir /dedoc_root
 RUN mkdir /dedoc_root/dedoc
-ADD dedoc/config.py /dedoc_root/dedoc/config.py
+ADD docker_gpu/config_on_gpu.py /dedoc_root/dedoc/config.py
 ADD dedoc/download_models.py /dedoc_root/dedoc/download_models.py
 RUN python3 /dedoc_root/dedoc/download_models.py
 

docker_gpu/config_on_gpu.py

@@ -0,0 +1,85 @@
+import importlib.util
+import logging
+import os
+import sys
+from typing import Any, Optional
+
+logging.basicConfig(stream=sys.stdout, level=logging.INFO, format="%(asctime)s - %(pathname)s - %(levelname)s - %(message)s")
+
+DEBUG_MODE = False
+RESOURCES_PATH = os.environ.get("RESOURCES_PATH", os.path.join(os.path.expanduser("~"), ".cache", "dedoc", "resources"))
+
+_config = dict(
+    # -----------------------------------------RESOURCES PATH SETTINGS----------------------------------------------------
+    resources_path=RESOURCES_PATH,
+    intermediate_data_path=os.path.join(RESOURCES_PATH, "datasets"),
+
+    # -----------------------------------------COMMON DEBUG SETTINGS----------------------------------------------------
+    debug_mode=DEBUG_MODE,
+    path_debug=os.path.join(os.path.abspath(os.sep), "tmp", "dedoc"),
+
+    # --------------------------------------------JOBLIB SETTINGS-------------------------------------------------------
+    # number of parallel jobs in some tasks as OCR
+    n_jobs=1,
+
+    # --------------------------------------------GPU SETTINGS----------------------------------------------------------
+    # set gpu in XGBoost and torch models
+    on_gpu=True,
+
+    # ---------------------------------------------API SETTINGS---------------------------------------------------------
+    # max file size in bytes
+    max_content_length=512 * 1024 * 1024,
+    # application port
+    api_port=int(os.environ.get("DOCREADER_PORT", "1231")),
+    static_files_dirs={},
+    # log settings
+    logger=logging.getLogger(),
+    import_path_init_api_args="dedoc.api.api_args",
+
+    # ----------------------------------------TABLE RECOGNIZER DEBUG SETTINGS-------------------------------------------
+    # path to save debug images for tables recognizer
+    path_detect=os.path.join(os.path.abspath(os.sep), "tmp", "dedoc", "debug_tables", "imgs", "detect_lines"),
+
+    # -------------------------------------------RECOGNIZE SETTINGS-----------------------------------------------------
+    # TESSERACT OCR confidence threshold ( values: [-1 - undefined;  0.0 : 100.0 % - confidence value)
+    ocr_conf_threshold=40.0,
+    # max depth of document structure tree
+    recursion_deep_subparagraphs=30
+)
+
+
+class Configuration(object):
+    """
+    Pattern Singleton for configuration service
+    INFO: Configuration class and config are created once at the first call
+    """
+    __instance = None
+    __config = None
+
+    @classmethod
+    def get_instance(cls: "Configuration") -> "Configuration":
+        """
+        Actual object creation will happen when we use Configuration.getInstance()
+        """
+        if not cls.__instance:
+            cls.__instance = Configuration()
+
+        return cls.__instance
+
+    def __init_config(self, args: Optional[Any] = None) -> None:
+        if args is not None and args.config_path is not None:
+            spec = importlib.util.spec_from_file_location("config_module", args.config_path)
+            config_module = importlib.util.module_from_spec(spec)
+            spec.loader.exec_module(config_module)
+            self.__config = config_module._config
+        else:
+            self.__config = _config
+
+    def get_config(self, args: Optional[Any] = None) -> dict:
+        if self.__config is None or args is not None:
+            self.__init_config(args)
+        return self.__config
+
+
+def get_config(args: Optional[Any] = None) -> dict:
+    return Configuration.get_instance().get_config(args)

docker_gpu/docker-compose.yml

@@ -4,8 +4,8 @@ services:
   dedoc:
     mem_limit: 16G
     build:
-      context: .
-      dockerfile: Dockerfile
+      context: ..
+      dockerfile: docker_gpu/Dockerfile
     restart: always
     tty: true
     ports:
@@ -19,8 +19,8 @@ services:
     depends_on:
       - dedoc
     build:
-          context: .
-          dockerfile: Dockerfile
+          context: ..
+          dockerfile: docker_gpu/Dockerfile
     tty: true
     environment:
       DOC_READER_HOST: "dedoc"