vlmparse 0.1.6__py3-none-any.whl → 0.1.8__py3-none-any.whl

vlmparse/servers/docker_server.py

@@ -2,15 +2,23 @@ import os
 from typing import Callable
 
 from loguru import logger
-from pydantic import BaseModel, Field
+from pydantic import Field
 
+from .model_identity import ModelIdentityMixin
 from .utils import docker_server
 
 
-class DockerServerConfig(BaseModel):
-    """Base configuration for deploying a Docker server."""
+class DockerServerConfig(ModelIdentityMixin):
+    """Base configuration for deploying a Docker server.
+
+    Inherits from ModelIdentityMixin which provides:
+    - model_name: str
+    - default_model_name: str | None
+    - aliases: list[str]
+    - _create_client_kwargs(base_url): Helper for creating client configs
+    - get_all_names(): All names this model can be referenced by
+    """
 
-    model_name: str
     docker_image: str
     dockerfile_dir: str | None = None
     command_args: list[str] = Field(default_factory=list)
@@ -27,7 +35,6 @@ class DockerServerConfig(BaseModel):
     environment: dict[str, str] = Field(default_factory=dict)
     volumes: dict[str, dict] | None = None
     entrypoint: str | None = None
-    aliases: list[str] = Field(default_factory=list)
 
     class Config:
         extra = "allow"
@@ -49,15 +56,14 @@ class DockerServerConfig(BaseModel):
 
     def update_command_args(
         self,
-        vllm_kwargs: dict | None = None,
-        forget_predefined_vllm_kwargs: bool = False,
+        vllm_args: dict | None = None,
+        forget_predefined_vllm_args: bool = False,
     ) -> list[str]:
-        if vllm_kwargs is not None:
-            new_kwargs = [f"--{k}={v}" for k, v in vllm_kwargs.items()]
-            if forget_predefined_vllm_kwargs:
-                self.command_args = new_kwargs
+        if vllm_args is not None:
+            if forget_predefined_vllm_args:
+                self.command_args = vllm_args
             else:
-                self.command_args.extend(new_kwargs)
+                self.command_args.extend(vllm_args)
 
         return self.command_args
 
@@ -85,22 +91,16 @@ class VLLMDockerServerConfig(DockerServerConfig):
     hf_home_folder: str | None = os.getenv("HF_HOME", None)
     add_model_key_to_server: bool = False
     container_port: int = 8000
-    aliases: list[str] = Field(default_factory=list)
-
-    @property
-    def llm_params(self):
-        from vlmparse.clients.openai_converter import LLMParams
-
-        return LLMParams(
-            base_url=f"http://localhost:{self.docker_port}{self.get_base_url_suffix()}",
-            model_name=self.default_model_name,
-        )
 
     @property
     def client_config(self):
         from vlmparse.clients.openai_converter import OpenAIConverterConfig
 
-        return OpenAIConverterConfig(llm_params=self.llm_params)
+        return OpenAIConverterConfig(
+            **self._create_client_kwargs(
+                f"http://localhost:{self.docker_port}{self.get_base_url_suffix()}"
+            )
+        )
 
     def get_command(self) -> list[str]:
         """Build VLLM-specific command."""
@@ -172,41 +172,64 @@ class ConverterServer:
     def stop(self):
         """Stop the Docker server."""
         if self._server_context is not None:
-            self._server_context.__exit__(None, None, None)
-            self._server_context = None
-            self._container = None
-            self.base_url = None
+            try:
+                self._server_context.__exit__(None, None, None)
+            except Exception as e:
+                logger.warning(f"Error during server cleanup: {e}")
+            finally:
+                self._server_context = None
+                self._container = None
+                self.base_url = None
             logger.info("Server stopped")
 
     def __del__(self):
-        """Automatically stop server when object is destroyed if auto_stop is True."""
-        if self.auto_stop and self._server_context is not None:
-            self.stop()
+        """Automatically stop server when object is destroyed if auto_stop is True.
+
+        Note: This is a fallback mechanism. Prefer using the context manager
+        or explicitly calling stop() for reliable cleanup.
+        """
+        try:
+            if self.auto_stop and self._server_context is not None:
+                self.stop()
+        except Exception:
+            pass  # Suppress errors during garbage collection
 
 
 class DockerConfigRegistry:
-    """Registry for mapping model names to their Docker configurations."""
+    """Registry for mapping model names to their Docker configurations.
+
+    Thread-safe registry that maps model names to their Docker configuration factories.
+    """
 
     def __init__(self):
-        self._registry = dict()
+        import threading
+
+        self._registry: dict[str, Callable[[], DockerServerConfig | None]] = {}
+        self._lock = threading.RLock()
 
     def register(
         self, model_name: str, config_factory: Callable[[], DockerServerConfig | None]
     ):
-        """Register a config factory for a model name."""
-        self._registry[model_name] = config_factory
+        """Register a config factory for a model name (thread-safe)."""
+        with self._lock:
+            self._registry[model_name] = config_factory
 
     def get(self, model_name: str, default=False) -> DockerServerConfig | None:
-        """Get config for a model name. Returns default if not registered."""
-        if model_name not in self._registry:
-            if default:
-                return VLLMDockerServerConfig(model_name=model_name)
-            return None
-        return self._registry[model_name]()
+        """Get config for a model name (thread-safe). Returns default if not registered."""
+        with self._lock:
+            if model_name not in self._registry:
+                if default:
+                    return VLLMDockerServerConfig(
+                        model_name=model_name, default_model_name=DEFAULT_MODEL_NAME
+                    )
+                return None
+            factory = self._registry[model_name]
+        return factory()
 
     def list_models(self) -> list[str]:
-        """List all registered model names."""
-        return list(self._registry.keys())
+        """List all registered model names (thread-safe)."""
+        with self._lock:
+            return list(self._registry.keys())
 
 
 # Global registry instance

vlmparse/servers/model_identity.py

@@ -0,0 +1,48 @@
+"""Model identity mixin for consistent model name handling between server and client configs."""
+
+from pydantic import BaseModel, Field
+
+
+class ModelIdentityMixin(BaseModel):
+    """Mixin providing model identity fields with validation.
+
+    This mixin ensures that model_name and default_model_name are consistently
+    passed from server configs to client configs.
+    """
+
+    model_name: str
+    default_model_name: str | None = None
+    aliases: list[str] = Field(default_factory=list)
+
+    def get_effective_model_name(self) -> str:
+        """Returns the model name to use for API calls."""
+        return self.default_model_name if self.default_model_name else self.model_name
+
+    def _create_client_kwargs(self, base_url: str) -> dict:
+        """Generate kwargs for client config with model identity.
+
+        Use this method in server configs to ensure consistent passing
+        of model_name and default_model_name to client configs.
+
+        Args:
+            base_url: The base URL for the client to connect to.
+
+        Returns:
+            Dictionary with base_url, model_name, and default_model_name.
+        """
+        return {
+            "base_url": base_url,
+            "model_name": self.model_name,
+            "default_model_name": self.get_effective_model_name(),
+        }
+
+    def get_all_names(self) -> list[str]:
+        """Get all names this model can be referenced by.
+
+        Returns:
+            List containing model_name, aliases, and short name (after last /).
+        """
+        names = [self.model_name] + self.aliases
+        if "/" in self.model_name:
+            names.append(self.model_name.split("/")[-1])
+        return [n for n in names if isinstance(n, str)]

vlmparse/servers/utils.py

@@ -2,6 +2,7 @@ import getpass
 import time
 from contextlib import contextmanager
 from pathlib import Path
+from urllib.parse import parse_qsl, urlparse
 
 import docker
 from loguru import logger
@@ -222,25 +223,52 @@ def docker_server(
             logger.info("Container stopped")
 
 
+def normalize_uri(uri: str) -> tuple:
+    u = urlparse(uri)
+
+    # --- Normalize scheme ---
+    scheme = (u.scheme or "http").lower()
+
+    # --- Normalize host ---
+    host = (u.hostname or "").lower()
+    if host in ("localhost", "0.0.0.0"):
+        host = "localhost"
+
+    # --- Normalize port (apply defaults) ---
+    if u.port:
+        port = u.port
+    else:
+        port = 443 if scheme == "https" else 80
+
+    # --- Normalize path ---
+    # Treat empty path as "/" and remove trailing slash (except root)
+    path = u.path or "/"
+    if path != "/" and path.endswith("/"):
+        path = path.rstrip("/")
+
+    # Collapse duplicate slashes
+    while "//" in path:
+        path = path.replace("//", "/")
+
+    # --- Normalize query parameters (sorted) ---
+    query_pairs = parse_qsl(u.query, keep_blank_values=True)
+    query = "&".join(f"{k}={v}" for k, v in sorted(query_pairs))
+
+    return (scheme, host, port, path, query)
+
+
 def get_model_from_uri(uri: str) -> str:
     model = None
     client = docker.from_env()
     containers = client.containers.list()
+
+    uri = normalize_uri(uri)
+
     for container in containers:
         c_uri = container.labels.get("vlmparse_uri")
         c_model = container.labels.get("vlmparse_model_name")
-        if c_uri is not None:
-            c_uri = c_uri.replace("localhost", "0.0.0.0")
-
-        # Check if user URI matches container URI (ignoring /v1 suffix if missing)
-        if c_uri and (
-            c_uri == uri or c_uri.startswith(uri.rstrip("/")) or uri.startswith(c_uri)
-        ):
-            # Update URI to the correct one from container (likely has /v1)
-            if len(c_uri) > len(uri.rstrip("/")):
-                logger.info(f"Updating URI from {uri} to {c_uri}")
-                uri = c_uri
 
+        if c_uri and uri == normalize_uri(c_uri):
             # Infer model if not provided
             if model is None and c_model:
                 logger.info(f"Inferred model {c_model} from container")

vlmparse/utils.py

@@ -19,8 +19,19 @@ def from_base64(base64_str: str):
     return Image.open(BytesIO(image_data))
 
 
-def get_file_paths(inputs: str | list[str]):
-    # Expand file paths from glob patterns
+def get_file_paths(inputs: str | list[str], raise_on_empty: bool = False) -> list[str]:
+    """Expand file paths from glob patterns.
+
+    Args:
+        inputs: A string or list of strings containing file paths, glob patterns, or directories.
+        raise_on_empty: If True, raise FileNotFoundError when no files are found.
+
+    Returns:
+        List of valid file paths.
+
+    Raises:
+        FileNotFoundError: If raise_on_empty is True and no files are found.
+    """
     file_paths = []
     if isinstance(inputs, str):
         inputs = [inputs]
@@ -36,6 +47,8 @@ def get_file_paths(inputs: str | list[str]):
     file_paths = [f for f in file_paths if os.path.exists(f) and os.path.isfile(f)]
 
     if not file_paths:
+        if raise_on_empty:
+            raise FileNotFoundError("No files found matching the input patterns")
         logger.error("No PDF files found matching the inputs patterns")
 
     return file_paths

{vlmparse-0.1.6.dist-info → vlmparse-0.1.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vlmparse
-Version: 0.1.6
+Version: 0.1.8
 Requires-Python: >=3.11.0
 Description-Content-Type: text/markdown
 License-File: LICENSE
@@ -54,6 +54,12 @@ Dynamic: license-file
 
 # vlmparse
 
+<div align="center">
+
+[\[📜 arXiv coming soon\]] | [[Dataset (🤗Hugging Face)]](https://huggingface.co/datasets/pulsia/fr-bench-pdf2md) | [[pypi]](https://pypi.org/project/vlmparse/) | [[vlmparse]](https://github.com/ld-lab-pulsia/vlmparse) | [[Benchmark]](https://github.com/ld-lab-pulsia/benchpdf2md)
+
+</div>
+
 A unified wrapper for Vision Language Models (VLM) and OCR solutions to parse PDF documents into Markdown.
 
 Features:
@@ -209,3 +215,7 @@ with ConverterWithServer(model="mineru2.5") as converter_with_server:
 ```
 
 Note that if you pass an uri of a vllm server to `ConverterWithServer`, the model name is inferred automatically and no server is started.
+
+## Credits
+
+This work was financed by La Poste and led by members of Probayes and OpenValue, two subsidiaries (filiales) of La Poste.

vlmparse-0.1.8.dist-info/RECORD

@@ -0,0 +1,38 @@
+vlmparse/base_model.py,sha256=4U4UPe8SNArliKnUf8pp8zQugWYsnhg9okylt7mrW1U,381
+vlmparse/build_doc.py,sha256=fb7awoqVN-6NBlKVkMFb1v1iTWcxne5QAyNaKYTyvM4,2275
+vlmparse/cli.py,sha256=asew0JdpbgFZrZqnG-Bqh5A_DrXcP0XomLB3y3AgG6Y,12855
+vlmparse/constants.py,sha256=DYaK7KtTW8p9MPb3iPvoP5H1r7ICRuIFo89P01q4uCI,184
+vlmparse/converter.py,sha256=KKcXqrp3nJo3d7DXjHn3O2SklbsJ489rDY4NJ9O42Fs,8795
+vlmparse/converter_with_server.py,sha256=A84l3YNal-hs2mMlER1sB29rddsO8MNOP2j9ts0ujtE,7280
+vlmparse/registries.py,sha256=B4kxibP7XYbhL9bZ5gn21LQCPhHCYftAM4i0-xD9fRs,6469
+vlmparse/utils.py,sha256=6Ff9OfAIVR-4_37QD5sifoNt_GmB3YUqgFwmIjuemtc,1727
+vlmparse/clients/chandra.py,sha256=zAHjgI_MJ5FVGANHCG8KJQByaw6-zTS6CHXsCBA8TJI,13025
+vlmparse/clients/deepseekocr.py,sha256=pKdNJD9v86BRn7YrXE6PGk_jQxnbZ_6UjgSUxgd3Su4,6859
+vlmparse/clients/docling.py,sha256=BLtNAxVJR6qvPip4ZBP-se8IMNFSbJ-fWEGlTSwimK8,5310
+vlmparse/clients/dotsocr.py,sha256=oAUzDMTObeW0sTy5sFl08O6GQPSTic5ITbJYh_45Z54,10414
+vlmparse/clients/granite_docling.py,sha256=KYaEdgk3oD0TuYDKqTQ4o6IkXC-E3AIYJ2KYxnEsjWM,3595
+vlmparse/clients/hunyuanocr.py,sha256=etpIiA28OoGW-o5pOGeBxOlUDjUQ4zcKXWnJ8ba44DU,1979
+vlmparse/clients/lightonocr.py,sha256=ZWC12U6myDr_2EuOPYGyJYxpBachjOUtLrxS62A8mzg,2048
+vlmparse/clients/mineru.py,sha256=9r6_JwCGB6-C-YKpoadUs7lSuDEJyBFkfjWV5WN-VrE,3599
+vlmparse/clients/mistral_converter.py,sha256=_hEyK_2vM5LEwbt30bFodMrWJtavLsBDxCkVclNUF4M,2837
+vlmparse/clients/nanonetocr.py,sha256=gTbD4OtuHiWd6Ack6Bx-anZM9P_aErfSHXwtymETvqM,1665
+vlmparse/clients/olmocr.py,sha256=V4638WftLCTr5Q6ZRgWKKSPAhFYdpBw3izeuda6EKDQ,1966
+vlmparse/clients/openai_converter.py,sha256=bckm33Pkvqul--DjfEEEI3evn4_va0CoQcigdpCCMGc,7746
+vlmparse/clients/paddleocrvl.py,sha256=q3AgEWj0UyXGpSEVZISdfqv2PV_qY-uF498bL8U1tpg,1596
+vlmparse/clients/prompts.py,sha256=-J60lqxgRzlkQ9VsQLxmWsIMaDt-gNqWqWoqHIw9CLc,4228
+vlmparse/clients/pipe_utils/cleaner.py,sha256=oxBkBTOkluN1lmeNbzajRIe0_D__ZGwUOBaI_Ph0uxE,2396
+vlmparse/clients/pipe_utils/html_to_md_conversion.py,sha256=cFFqzD2jCNw_968_eu3Wt--Ox7iJj2Rn5UoP_DZWosU,4112
+vlmparse/clients/pipe_utils/utils.py,sha256=935ecIO446I0pstszE_1nrIPHn1Ffrxunq7fVd0dsd8,315
+vlmparse/data_model/box.py,sha256=lJsh4qhjgYXZF5vTSJ1qMXD5GVlBi2_SBedBMlfJikU,16868
+vlmparse/data_model/document.py,sha256=xheaMeStOj2c9GZKmdtxcEl_Dj44V5JyVp6JnTrSpH0,4615
+vlmparse/servers/docker_server.py,sha256=FOIHU0_CDfyZ9UA285BrnUFuEMJRxbu-OzlByBa-P9s,7951
+vlmparse/servers/model_identity.py,sha256=DkH7KQAAZA9Sn7eJEnaKfH54XSEI17aqD1ScqqkTBEk,1711
+vlmparse/servers/utils.py,sha256=tIXhgbF9EVOJy2nYEguVq69gn9ATxtya_1F4wZSt68o,9454
+vlmparse/st_viewer/fs_nav.py,sha256=7GNH68h2Loh5pQ64Pe72-D2cs2BLhqRXevEmKdFmPX0,1616
+vlmparse/st_viewer/st_viewer.py,sha256=m2rQTtk5rlwErNmivNAg-4rkHkvNkvLhoJZxFQi7Dwk,2105
+vlmparse-0.1.8.dist-info/licenses/LICENSE,sha256=3TKJHk8hPBR5dbLWZ3IpfCftl-_m-iyBwpYQGZYxj14,1080
+vlmparse-0.1.8.dist-info/METADATA,sha256=dwu5tiTLuhVMYL-ZQCMNYW_MNlJu84V2us0aeRfrSpU,6048
+vlmparse-0.1.8.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+vlmparse-0.1.8.dist-info/entry_points.txt,sha256=gD5berP6HwE2wNIkls-Lw5goiceA8uMgPEd7ifnFJXs,47
+vlmparse-0.1.8.dist-info/top_level.txt,sha256=k4ni-GNH_iAX7liQEsk_KY_c3xgZgt8k9fsSs9IXLXs,9
+vlmparse-0.1.8.dist-info/RECORD,,

{vlmparse-0.1.6.dist-info → vlmparse-0.1.8.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.9.0)
+Generator: setuptools (80.10.2)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

vlmparse-0.1.6.dist-info/RECORD

@@ -1,36 +0,0 @@
-vlmparse/base_model.py,sha256=4U4UPe8SNArliKnUf8pp8zQugWYsnhg9okylt7mrW1U,381
-vlmparse/build_doc.py,sha256=LAWrnFrqamN5PwJo57AUtQOPrMFGnCGw4gBjEKZ6pYo,2127
-vlmparse/cli.py,sha256=gY45YZe5LanN-ozG2vVtOaB2qyNWpjO2DvPmJeBi_wA,13045
-vlmparse/constants.py,sha256=7-47S01n4MI2ebR09bpdOo3_P16d-z-NVGsm6KJP8ls,110
-vlmparse/converter.py,sha256=F0JSY9sFYUggCvaUCb27kKGJJpnZKW2FStMDVJoIOeQ,7383
-vlmparse/converter_with_server.py,sha256=62kcEp0NjzDR2vVmEfCeeLlwbb8E3sWcseb2jjK7DpM,4861
-vlmparse/registries.py,sha256=yBVrrhy61rSoLwdNV-z0C4lqIpTbLoWab3V6u7aSyNM,5797
-vlmparse/utils.py,sha256=rcVrtPiQVj_8HAmFQOu___72uYIapp_X89yxrMNCBow,1236
-vlmparse/clients/chandra.py,sha256=EulsCZdwOtm0pQ6CDm320U96k8aWFN4wKqCm1Xo7VCE,9775
-vlmparse/clients/deepseekocr.py,sha256=Uw6tPvP2KVsPDlz1ZUgYdbgQSjmFPuYeFDrGMMOTBAo,6501
-vlmparse/clients/docling.py,sha256=SAkLsqseuWfkuiel8FWR1G0Z5s-SZU3dE2JbsOvF4SA,5328
-vlmparse/clients/dotsocr.py,sha256=uGJoYEiDkP3-rmfdkAnMeAX-T4RZyEPoh6jmow5_-J8,10336
-vlmparse/clients/granite_docling.py,sha256=EQpsv5qSJG0HtMSacmJStER2sq4TGf1EMU5_NmJsl4g,4634
-vlmparse/clients/hunyuanocr.py,sha256=UFqaS4b8UM9EtizyrZIxlqcYlESmxm8xrQZP7lL6tkE,1857
-vlmparse/clients/lightonocr.py,sha256=wx1Im8Z3wlRWwYbPqnSd3LqTtdAU8CnX5mzu1BuCUY8,1314
-vlmparse/clients/mineru.py,sha256=6jZ1sKn2kGwUvD8gVs4PqEDH7uUXYK8pAB5Fr1JeqnY,3617
-vlmparse/clients/nanonetocr.py,sha256=BT5vaeerCsK5agvOaHK3NvLUqWd1FfDmrMmDYbp646I,1543
-vlmparse/clients/olmocr.py,sha256=A4Vl0meYpU5QPTML_OxyyRM07xCxtfrMZedgGMYEcuU,1851
-vlmparse/clients/openai_converter.py,sha256=nMKJeWH43UxHMMLns3wjX0pYjU5Xnai6IYxFmS9I63s,6193
-vlmparse/clients/paddleocrvl.py,sha256=qFBDj_UQocyq3WCh24tUOx9Ud7S9DfSm-1n3ztikY2s,1402
-vlmparse/clients/prompts.py,sha256=-J60lqxgRzlkQ9VsQLxmWsIMaDt-gNqWqWoqHIw9CLc,4228
-vlmparse/clients/pipe_utils/cleaner.py,sha256=oxBkBTOkluN1lmeNbzajRIe0_D__ZGwUOBaI_Ph0uxE,2396
-vlmparse/clients/pipe_utils/html_to_md_conversion.py,sha256=cFFqzD2jCNw_968_eu3Wt--Ox7iJj2Rn5UoP_DZWosU,4112
-vlmparse/clients/pipe_utils/utils.py,sha256=935ecIO446I0pstszE_1nrIPHn1Ffrxunq7fVd0dsd8,315
-vlmparse/data_model/box.py,sha256=lJsh4qhjgYXZF5vTSJ1qMXD5GVlBi2_SBedBMlfJikU,16868
-vlmparse/data_model/document.py,sha256=xheaMeStOj2c9GZKmdtxcEl_Dj44V5JyVp6JnTrSpH0,4615
-vlmparse/servers/docker_server.py,sha256=FBW2TvtUHUQNwj0rBP92shvoiJCGlc_oAKQBXN8260E,7114
-vlmparse/servers/utils.py,sha256=qy2-rnQTCQKt6CeTV5H74tvRTXyzBV2KswQiYW8Tf-k,8908
-vlmparse/st_viewer/fs_nav.py,sha256=7GNH68h2Loh5pQ64Pe72-D2cs2BLhqRXevEmKdFmPX0,1616
-vlmparse/st_viewer/st_viewer.py,sha256=m2rQTtk5rlwErNmivNAg-4rkHkvNkvLhoJZxFQi7Dwk,2105
-vlmparse-0.1.6.dist-info/licenses/LICENSE,sha256=3TKJHk8hPBR5dbLWZ3IpfCftl-_m-iyBwpYQGZYxj14,1080
-vlmparse-0.1.6.dist-info/METADATA,sha256=Xad3SjAYvCzUvPo6A6GKvc3daxtf5XNs1AQjDlF7RmI,5597
-vlmparse-0.1.6.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-vlmparse-0.1.6.dist-info/entry_points.txt,sha256=gD5berP6HwE2wNIkls-Lw5goiceA8uMgPEd7ifnFJXs,47
-vlmparse-0.1.6.dist-info/top_level.txt,sha256=k4ni-GNH_iAX7liQEsk_KY_c3xgZgt8k9fsSs9IXLXs,9
-vlmparse-0.1.6.dist-info/RECORD,,