chore: Add verbose option to ExtractionStrategy classes

This commit adds a new `verbose` option to the `ExtractionStrategy` classes. The `verbose` option allows for logging of extraction details, such as the number of extracted blocks and the URL being processed. This improves the debugging and monitoring capabilities of the code.

crawl4ai/extraction_strategy.py

@@ -19,6 +19,7 @@ class ExtractionStrategy(ABC):
     def __init__(self, **kwargs):
         self.DEL = "<|DEL|>"
         self.name = self.__class__.__name__
+        self.verbose = kwargs.get("verbose", False)
 
     @abstractmethod
     def extract(self, url: str, html: str, *q, **kwargs) -> List[Dict[str, Any]]:
@@ -65,6 +66,7 @@ class LLMExtractionStrategy(ExtractionStrategy):
         self.provider = provider
         self.api_token = api_token or PROVIDER_MODELS.get(provider, None) or os.getenv("OPENAI_API_KEY")
         self.instruction = instruction
+        self.verbose = kwargs.get("verbose", False)
         
         if not self.api_token:
             raise ValueError("API token must be provided for LLMExtractionStrategy. Update the config.py or set OPENAI_API_KEY environment variable.")
@@ -105,6 +107,7 @@ class LLMExtractionStrategy(ExtractionStrategy):
                     "content": unparsed
                 })
         
+        if self.verbose:
             print("[LOG] Extracted", len(blocks), "blocks from URL:", url, "block index:", ix)
         return blocks
     
@@ -171,6 +174,7 @@ class CosineStrategy(ExtractionStrategy):
         self.linkage_method = linkage_method
         self.top_k = top_k
         self.timer = time.time()
+        self.verbose = kwargs.get("verbose", False)
         
         self.buffer_embeddings = np.array([])
 
@@ -180,6 +184,8 @@ class CosineStrategy(ExtractionStrategy):
             self.tokenizer, self.model = load_bge_small_en_v1_5()
 
         self.nlp = load_text_multilabel_classifier()
+        
+        if self.verbose:
             print(f"[LOG] Model loaded {model_name}, models/reuters, took " + str(time.time() - self.timer) + " seconds")
 
     def filter_documents_embeddings(self, documents: List[str], semantic_filter: str, threshold: float = 0.5) -> List[str]:

crawl4ai/model_loader.py

@@ -15,15 +15,15 @@ def get_home_folder():
 @lru_cache()
 def load_bert_base_uncased():
     from transformers import BertTokenizer, BertModel, AutoTokenizer, AutoModel
-    tokenizer = BertTokenizer.from_pretrained('bert-base-uncased') #, resume_download=None)
-    model = BertModel.from_pretrained('bert-base-uncased') #, resume_download=None)
+    tokenizer = BertTokenizer.from_pretrained('bert-base-uncased', resume_download=None)
+    model = BertModel.from_pretrained('bert-base-uncased', resume_download=None)
     return tokenizer, model
 
 @lru_cache()
 def load_bge_small_en_v1_5():
     from transformers import BertTokenizer, BertModel, AutoTokenizer, AutoModel
-    tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-small-en-v1.5') #, resume_download=None)
-    model = AutoModel.from_pretrained('BAAI/bge-small-en-v1.5') #, resume_download=None)
+    tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-small-en-v1.5', resume_download=None)
+    model = AutoModel.from_pretrained('BAAI/bge-small-en-v1.5', resume_download=None)
     model.eval()
     return tokenizer, model
 

crawl4ai/web_crawler.py

@@ -86,6 +86,7 @@ class WebCrawler:
         **kwargs,
     ) -> CrawlResult:
         extraction_strategy = extraction_strategy or NoExtractionStrategy()
+        extraction_strategy.verbose = verbose
         # Check if extraction strategy is an instance of ExtractionStrategy if not raise an error
         if not isinstance(extraction_strategy, ExtractionStrategy):
             raise ValueError("Unsupported extraction strategy")