vlmparse 0.1.9__py3-none-any.whl → 0.1.10__py3-none-any.whl

vlmparse/cli.py

@@ -15,8 +15,8 @@ def serve(
         None,
         help='Comma-separated GPU device IDs (e.g., "0" or "0,1,2"). If not specified, all GPUs will be used.',
     ),
-    server: Literal["registry", "hf"] = typer.Option(
-        "registry", help="Server type for the model. 'registry' (default) or 'hf'."
+    provider: Literal["registry", "hf"] = typer.Option(
+        "registry", help="provider type for the model. 'registry' (default) or 'hf'."
     ),
     vllm_args: list[str] | None = typer.Option(
         None,
@@ -48,7 +48,7 @@ def serve(
         model=model,
         gpus=gpus,
         port=port,
-        server=server,
+        provider=provider,
         vllm_args=vllm_args,
         forget_predefined_vllm_args=forget_predefined_vllm_args,
         auto_stop=False,
@@ -99,7 +99,7 @@ def convert(
             "image_description (describe the image), formula (formula extraction), chart (chart recognition)"
         ),
     ),
-    server: Literal["registry", "hf", "google", "openai"] = typer.Option(
+    provider: Literal["registry", "hf", "google", "openai"] = typer.Option(
         "registry",
         help="Server type for the model. Defaults to 'registry'.",
     ),
@@ -127,21 +127,20 @@ def convert(
         gpus: Comma-separated GPU device IDs (e.g., "0" or "0,1,2"). If not specified, all GPUs will be used.
         mode: Output mode - "document" (save as JSON zip), "md" (save as markdown file), "md_page" (save as folder of markdown pages)
         conversion_mode: Conversion mode - "ocr" (plain), "ocr_layout" (OCR with layout), "table" (table-centric), "image_description" (describe the image), "formula" (formula extraction), "chart" (chart recognition)
-        server: Server type for the model. Defaults to 'registry'.
-        with_vllm_server: Deprecated. Use --server hf instead. If True, a local VLLM server will be deployed if the model is not found in the registry. Note that if the model is in the registry and the uri is None, the server will be anyway deployed.
+        provider: provider type for the model. Defaults to 'registry'.
         dpi: DPI to use for the conversion. If not specified, the default DPI will be used.
         debug: If True, run in debug mode (single-threaded, no concurrency)
     """
     from vlmparse.converter_with_server import ConverterWithServer
 
-    if with_vllm_server and server == "registry":
-        server = "hf"
+    if with_vllm_server and provider == "registry":
+        provider = "hf"
 
     with ConverterWithServer(
         model=model,
         uri=uri,
         gpus=gpus,
-        server=server,
+        provider=provider,
         concurrency=concurrency,
         return_documents=_return_documents,
     ) as converter_with_server:

vlmparse/converter_with_server.py

@@ -14,7 +14,7 @@ def start_server(
     model: str,
     gpus: str,
     port: None | int = None,
-    server: Literal["registry", "hf"] = "registry",
+    provider: Literal["registry", "hf"] = "registry",
     vllm_args: list[str] = {},
     forget_predefined_vllm_args: bool = False,
     auto_stop: bool = False,
@@ -33,12 +33,12 @@ def start_server(
         port = DEFAULT_SERVER_PORT
 
     if docker_config is None:
-        if server == "registry":
+        if provider == "registry":
             print(f"DEBUG: Registry lookup failed for {model} (strict mode)")
             raise ValueError(
-                f"Model '{model}' not found in registry and server='registry'. Use server='hf' to serve arbitrary HuggingFace models."
+                f"Model '{model}' not found in registry and provider='registry'. Use provider='hf' to serve arbitrary HuggingFace models."
             )
-        elif server == "hf":
+        elif provider == "hf":
             docker_config = VLLMDockerServerConfig(
                 model_name=model, default_model_name=DEFAULT_MODEL_NAME
             )
@@ -65,14 +65,14 @@ def start_server(
         logger.info(
             f"Deploying server for {docker_config.model_name} on port {port}..."
         )
-        server = docker_config.get_server(auto_stop=auto_stop)
-        if server is None:
+        provider = docker_config.get_server(auto_stop=auto_stop)
+        if provider is None:
             logger.error(f"Model server not found for model: {model}")
             return "", container, None, docker_config
 
-        base_url, container = server.start()
+        base_url, container = provider.start()
 
-    return base_url, container, server, docker_config
+    return base_url, container, provider, docker_config
 
 
 class ConverterWithServer:
@@ -82,7 +82,7 @@ class ConverterWithServer:
         uri: str | None = None,
         gpus: str | None = None,
         port: int | None = None,
-        server: Literal["registry", "hf", "google", "openai"] = "registry",
+        provider: Literal["registry", "hf", "google", "openai"] = "registry",
         concurrency: int = 10,
         vllm_args: dict | None = None,
         forget_predefined_vllm_args: bool = False,
@@ -98,7 +98,7 @@ class ConverterWithServer:
         self.uri = uri
         self.port = port
         self.gpus = gpus
-        self.server_type = server
+        self.provider = provider
         self.concurrency = concurrency
         self.vllm_args = vllm_args
         self.forget_predefined_vllm_args = forget_predefined_vllm_args
@@ -118,19 +118,19 @@ class ConverterWithServer:
 
         start_local_server = False
         if self.uri is None:
-            if self.server_type == "hf":
+            if self.provider == "hf":
                 start_local_server = True
-            elif self.server_type == "registry":
+            elif self.provider == "registry":
                 if self.model in docker_config_registry.list_models():
                     start_local_server = True
 
         if start_local_server:
-            server_arg = "hf" if self.server_type == "hf" else "registry"
+            server_arg = "hf" if self.provider == "hf" else "registry"
             _, _, self.server, docker_config = start_server(
                 model=self.model,
                 gpus=self.gpus,
                 port=self.port,
-                server=server_arg,
+                provider=server_arg,
                 vllm_args=self.vllm_args,
                 forget_predefined_vllm_args=self.forget_predefined_vllm_args,
                 auto_stop=True,
@@ -146,7 +146,7 @@ class ConverterWithServer:
                     return_documents_in_batch_mode=self.return_documents
                 )
 
-        elif self.server_type == "hf":
+        elif self.provider == "hf":
             client_config = OpenAIConverterConfig(
                 model_name=self.model, base_url=self.uri
             )

{vlmparse-0.1.9.dist-info → vlmparse-0.1.10.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vlmparse
-Version: 0.1.9
+Version: 0.1.10
 Requires-Python: >=3.11.0
 Description-Content-Type: text/markdown
 License-File: LICENSE
@@ -56,7 +56,7 @@ Dynamic: license-file
 
 <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)
+[\[📜 arXiv coming soon\]] | [[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) | [[Leaderboard]](https://huggingface.co/spaces/pulsia/fr-bench-pdf2md)
 
 </div>
 
@@ -71,7 +71,7 @@ Features:
 
 Supported Converters:
 
-- **Open Source Small VLMs**: `lightonocr`, `mineru2.5`, `hunyuanocr`, `paddleocrvl`, `granite-docling`, `olmocr2-fp8`, `dotsocr`, `chandra`, `deepseekocr`, `nanonets/Nanonets-OCR2-3B`
+- **Open Source Small VLMs**: `lightonocr2`, `mineru2.5`, `hunyuanocr`, `paddleocrvl-1.5`, `granite-docling`, `olmocr2-fp8`, `dotsocr`, `chandra`, `deepseekocr2`, `nanonets/Nanonets-OCR2-3B`
 - **Open Source Generalist VLMs**: such as the Qwen family.
 - **Pipelines**: `docling`
 - **Proprietary LLMs**: `gemini`, `gpt`
@@ -115,13 +115,13 @@ Note that you can bypass the previous installation step and just add uvx before
 With a general VLM (requires setting your api key as an environment variable):
 
 ```bash
-vlmparse convert --input "*.pdf" --out_folder ./output --model gemini-2.5-flash-lite
+vlmparse convert "*.pdf" --out_folder ./output --model gemini-2.5-flash-lite
 ```
 
 Convert with auto deployment of a small vlm (or any huggingface VLM model, requires a gpu + docker installation):
 
 ```bash
-vlmparse convert --input "*.pdf" --out_folder ./output --model nanonets/Nanonets-OCR2-3B
+vlmparse convert "*.pdf" --out_folder ./output --model nanonets/Nanonets-OCR2-3B
 ```
 
 ### Deploy a local model server
@@ -131,7 +131,7 @@ Deployment (requires a gpu + docker installation):
 - Check that the port is not used by another service.
 
 ```bash
-vlmparse serve --model lightonocr2 --port 8000 --gpus 1
+vlmparse serve lightonocr2 --port 8000 --gpus 1
 ```
 
 then convert:

{vlmparse-0.1.9.dist-info → vlmparse-0.1.10.dist-info}/RECORD

@@ -1,9 +1,9 @@
 vlmparse/base_model.py,sha256=4U4UPe8SNArliKnUf8pp8zQugWYsnhg9okylt7mrW1U,381
 vlmparse/build_doc.py,sha256=fb7awoqVN-6NBlKVkMFb1v1iTWcxne5QAyNaKYTyvM4,2275
-vlmparse/cli.py,sha256=CAEa6_BGrHxagR_T25pV4UIDY2u6Ep4p30pAk3SlelM,18514
+vlmparse/cli.py,sha256=jP_BnFaeW1rm3iTcdw5WFRfQUgDYd6HC1Zh-5JbE9_4,18285
 vlmparse/constants.py,sha256=DYaK7KtTW8p9MPb3iPvoP5H1r7ICRuIFo89P01q4uCI,184
 vlmparse/converter.py,sha256=KKcXqrp3nJo3d7DXjHn3O2SklbsJ489rDY4NJ9O42Fs,8795
-vlmparse/converter_with_server.py,sha256=Ol5llwOdYcsRIRHR1A1g0nL4pneYi4MzWtBld2LlP80,8863
+vlmparse/converter_with_server.py,sha256=nDGF-FEqskAECam_Sm8GbPMGPdI2Iua4lHaHbpMZx_k,8872
 vlmparse/registries.py,sha256=4xiDKyIzAW68ZWyOtUmBOvzcXVqTPPdeoxD2s9RbjZ0,6714
 vlmparse/utils.py,sha256=6Ff9OfAIVR-4_37QD5sifoNt_GmB3YUqgFwmIjuemtc,1727
 vlmparse/clients/chandra.py,sha256=zAHjgI_MJ5FVGANHCG8KJQByaw6-zTS6CHXsCBA8TJI,13025
@@ -36,9 +36,9 @@ vlmparse/servers/server_registry.py,sha256=FUF_XnN8872vKnc8-TrEBntwBS5i3ZYVJvTHr
 vlmparse/servers/utils.py,sha256=rbqn9i6XB1YOEFluP4Ur0Ma40_6riUxJ1eMS8LSWbKs,3998
 vlmparse/st_viewer/fs_nav.py,sha256=7GNH68h2Loh5pQ64Pe72-D2cs2BLhqRXevEmKdFmPX0,1616
 vlmparse/st_viewer/st_viewer.py,sha256=wg0qfhAKdvnkpc3xDK8QnWP9adjEThzeS-I5vHGDhIU,2132
-vlmparse-0.1.9.dist-info/licenses/LICENSE,sha256=3TKJHk8hPBR5dbLWZ3IpfCftl-_m-iyBwpYQGZYxj14,1080
-vlmparse-0.1.9.dist-info/METADATA,sha256=UNniHfumm45awE3CpNS7AT2mEDC28H_3H8KbnNoCf44,6022
-vlmparse-0.1.9.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-vlmparse-0.1.9.dist-info/entry_points.txt,sha256=gD5berP6HwE2wNIkls-Lw5goiceA8uMgPEd7ifnFJXs,47
-vlmparse-0.1.9.dist-info/top_level.txt,sha256=k4ni-GNH_iAX7liQEsk_KY_c3xgZgt8k9fsSs9IXLXs,9
-vlmparse-0.1.9.dist-info/RECORD,,
+vlmparse-0.1.10.dist-info/licenses/LICENSE,sha256=3TKJHk8hPBR5dbLWZ3IpfCftl-_m-iyBwpYQGZYxj14,1080
+vlmparse-0.1.10.dist-info/METADATA,sha256=OIRlJUlRioNzrehJIK2dmBcTFHI7A6H5uedu-EzDTQA,6077
+vlmparse-0.1.10.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+vlmparse-0.1.10.dist-info/entry_points.txt,sha256=gD5berP6HwE2wNIkls-Lw5goiceA8uMgPEd7ifnFJXs,47
+vlmparse-0.1.10.dist-info/top_level.txt,sha256=k4ni-GNH_iAX7liQEsk_KY_c3xgZgt8k9fsSs9IXLXs,9
+vlmparse-0.1.10.dist-info/RECORD,,