sparrow-parse 0.4.6__tar.gz → 0.4.8__tar.gz

{sparrow-parse-0.4.6 → sparrow-parse-0.4.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sparrow-parse
-Version: 0.4.6
+Version: 0.4.8
 Summary: Sparrow Parse is a Python package (part of Sparrow) for parsing and extracting information from documents.
 Home-page: https://github.com/katanaml/sparrow/tree/main/sparrow-data/parse
 Author: Andrej Baranovskij

{sparrow-parse-0.4.6 → sparrow-parse-0.4.8}/setup.py

@@ -8,7 +8,7 @@ with open("requirements.txt", "r", encoding="utf-8") as fh:
 
 setup(
     name="sparrow-parse",
-    version="0.4.6",
+    version="0.4.8",
     author="Andrej Baranovskij",
     author_email="andrejus.baranovskis@gmail.com",
     description="Sparrow Parse is a Python package (part of Sparrow) for parsing and extracting information from documents.",

sparrow-parse-0.4.8/sparrow_parse/__init__.py

@@ -0,0 +1 @@
+__version__ = '0.4.8'

{sparrow-parse-0.4.6 → sparrow-parse-0.4.8}/sparrow_parse/extractors/vllm_extractor.py

@@ -152,34 +152,34 @@ if __name__ == "__main__":
 
     extractor = VLLMExtractor()
 
-    # export HF_TOKEN="hf_"
-    config = {
-        "method": "mlx",  # Could be 'huggingface', 'mlx' or 'local_gpu'
-        "model_name": "mlx-community/Qwen2-VL-7B-Instruct-8bit",
-        # "hf_space": "katanaml/sparrow-qwen2-vl-7b",
-        # "hf_token": os.getenv('HF_TOKEN'),
-        # Additional fields for local GPU inference
-        # "device": "cuda", "model_path": "model.pth"
-    }
-
-    # Use the factory to get the correct instance
-    factory = InferenceFactory(config)
-    model_inference_instance = factory.get_inference_instance()
-
-    input_data = [
-        {
-            "file_path": "/Users/andrejb/Work/katana-git/sparrow/sparrow-ml/llm/data/invoice_1.jpg",
-            "text_input": "retrieve document data. return response in JSON format"
-        }
-    ]
-
-    # Now you can run inference without knowing which implementation is used
-    results_array, num_pages = extractor.run_inference(model_inference_instance, input_data, tables_only=True,
-                                                       generic_query=False,
-                                                       debug_dir="/Users/andrejb/Work/katana-git/sparrow/sparrow-ml/llm/data/",
-                                                       debug=True,
-                                                       mode=None)
-
-    for i, result in enumerate(results_array):
-        print(f"Result for page {i + 1}:", result)
-    print(f"Number of pages: {num_pages}")
+    # # export HF_TOKEN="hf_"
+    # config = {
+    #     "method": "mlx",  # Could be 'huggingface', 'mlx' or 'local_gpu'
+    #     "model_name": "mlx-community/Qwen2-VL-7B-Instruct-8bit",
+    #     # "hf_space": "katanaml/sparrow-qwen2-vl-7b",
+    #     # "hf_token": os.getenv('HF_TOKEN'),
+    #     # Additional fields for local GPU inference
+    #     # "device": "cuda", "model_path": "model.pth"
+    # }
+    #
+    # # Use the factory to get the correct instance
+    # factory = InferenceFactory(config)
+    # model_inference_instance = factory.get_inference_instance()
+    #
+    # input_data = [
+    #     {
+    #         "file_path": "/Users/andrejb/Work/katana-git/sparrow/sparrow-ml/llm/data/invoice_1.jpg",
+    #         "text_input": "retrieve document data. return response in JSON format"
+    #     }
+    # ]
+    #
+    # # Now you can run inference without knowing which implementation is used
+    # results_array, num_pages = extractor.run_inference(model_inference_instance, input_data, tables_only=True,
+    #                                                    generic_query=False,
+    #                                                    debug_dir="/Users/andrejb/Work/katana-git/sparrow/sparrow-ml/llm/data/",
+    #                                                    debug=True,
+    #                                                    mode=None)
+    #
+    # for i, result in enumerate(results_array):
+    #     print(f"Result for page {i + 1}:", result)
+    # print(f"Number of pages: {num_pages}")

{sparrow-parse-0.4.6 → sparrow-parse-0.4.8}/sparrow_parse/vllm/inference_factory.py

@@ -23,9 +23,3 @@ class InferenceFactory:
         # model = torch.load('model.pth')
         # return model
         raise NotImplementedError("Model loading logic not implemented")
-
-
-    def unload_inference_instance(self, instance):
-        if instance and hasattr(instance, "unload_model"):
-            instance.unload_model()
-            print(f"Inference instance of type {type(instance).__name__} has been unloaded.")

{sparrow-parse-0.4.6 → sparrow-parse-0.4.8}/sparrow_parse/vllm/mlx_inference.py

@@ -4,7 +4,7 @@ from mlx_vlm.utils import load_image
 from sparrow_parse.vllm.inference_base import ModelInference
 import os
 import json
-import gc
+from rich import print
 
 
 class MLXInference(ModelInference):
@@ -20,40 +20,19 @@ class MLXInference(ModelInference):
         :param model_name: Name of the model to load.
         """
         self.model_name = model_name
-        self.model = None
-        self.processor = None
-        print(f"MLXInference initialized with model: {model_name}")
+        print(f"MLXInference initialized for model: {model_name}")
 
 
-    def unload_model(self):
-        """
-        Unload the model and release resources.
-        """
-        if self.model:
-            print(f"Unloading model: {self.model_name}")
-            del self.model
-            self.model = None
-        if self.processor:
-            print(f"Unloading processor for model: {self.model_name}")
-            del self.processor
-            self.processor = None
-
-        # Force garbage collection to release memory
-        gc.collect()
-        print(f"Model {self.model_name} and its resources have been unloaded, memory cleared.")
-
-
-    def _load_model_and_processor(self, model_name):
+    @staticmethod
+    def _load_model_and_processor(model_name):
         """
         Load the model and processor for inference.
+
         :param model_name: Name of the model to load.
         :return: Tuple containing the loaded model and processor.
         """
-        print(f"Loading model and processor for: {model_name}...")
         model, processor = load(model_name)
-        self.model = model  # Store model instance
-        self.processor = processor  # Store processor instance
-        print(f"Model and processor for '{model_name}' loaded successfully.")
+        print(f"Loaded model: {model_name}")
         return model, processor
 
 
@@ -103,54 +82,48 @@ class MLXInference(ModelInference):
     def inference(self, input_data, mode=None):
         """
         Perform inference on input data using the specified model.
+
         :param input_data: A list of dictionaries containing image file paths and text inputs.
         :param mode: Optional mode for inference ("static" for simple JSON output).
         :return: List of processed model responses.
         """
-        try:
-            if mode == "static":
-                return [self.get_simple_json()]
-
-            # Load the model and processor
-            model, processor = self._load_model_and_processor(self.model_name)
-            config = model.config
-
-            # Prepare absolute file paths
-            file_paths = self._extract_file_paths(input_data)
-
-            results = []
-            for file_path in file_paths:
-                image, width, height = self.load_image_data(file_path)
-
-                # Prepare messages for the chat model
-                messages = [
-                    {"role": "system",
-                     "content": "You are an expert at extracting structured text from image documents."},
-                    {"role": "user", "content": input_data[0]["text_input"]},
-                ]
-
-                # Generate and process response
-                prompt = apply_chat_template(processor, config, messages)
-                response = generate(
-                    model,
-                    processor,
-                    image,
-                    prompt,
-                    resize_shape=(width, height),
-                    max_tokens=4000,
-                    temperature=0.0,
-                    verbose=False
-                )
-                results.append(self.process_response(response))
-
-                print("Inference completed successfully for: ", file_path)
-
-            return results
-
-        finally:
-            # Always unload the model after inference
-            self.unload_model()
+        if mode == "static":
+            return [self.get_simple_json()]
+
+        # Load the model and processor
+        model, processor = self._load_model_and_processor(self.model_name)
+        config = model.config
+
+        # Prepare absolute file paths
+        file_paths = self._extract_file_paths(input_data)
+
+        results = []
+        for file_path in file_paths:
+            image, width, height = self.load_image_data(file_path)
+
+            # Prepare messages for the chat model
+            messages = [
+                {"role": "system", "content": "You are an expert at extracting structured text from image documents."},
+                {"role": "user", "content": input_data[0]["text_input"]},
+            ]
+
+            # Generate and process response
+            prompt = apply_chat_template(processor, config, messages)  # Assuming defined
+            response = generate(
+                model,
+                processor,
+                image,
+                prompt,
+                resize_shape=(width, height),
+                max_tokens=4000,
+                temperature=0.0,
+                verbose=False
+            )
+            results.append(self.process_response(response))
+
+            print("Inference completed successfully for: ", file_path)
 
+        return results
 
     @staticmethod
     def _extract_file_paths(input_data):

{sparrow-parse-0.4.6 → sparrow-parse-0.4.8}/sparrow_parse.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sparrow-parse
-Version: 0.4.6
+Version: 0.4.8
 Summary: Sparrow Parse is a Python package (part of Sparrow) for parsing and extracting information from documents.
 Home-page: https://github.com/katanaml/sparrow/tree/main/sparrow-data/parse
 Author: Andrej Baranovskij

sparrow-parse-0.4.6/sparrow_parse/__init__.py

@@ -1 +0,0 @@
-__version__ = '0.4.6'