langroid 0.48.0__py3-none-any.whl → 0.48.1__py3-none-any.whl

langroid/mytypes.py

@@ -45,6 +45,8 @@ class DocMetaData(BaseModel):
 
     source: str = "context"  # just reference
     source_content: str = "context"  # reference and content
+    title: str = "unknown"
+    published_date: str = "unknown"
     is_chunk: bool = False  # if it is a chunk, don't split
     id: str = Field(default_factory=lambda: str(uuid4()))
     window_ids: List[str] = []  # for RAG: ids of chunks around this one

langroid/parsing/url_loader.py

@@ -258,7 +258,13 @@ class FirecrawlCrawler(BaseCrawler):
                     with open(filename, "w") as f:
                         f.write(content)
                     docs.append(
-                        Document(content=content, metadata=DocMetaData(source=url))
+                        Document(
+                            content=content,
+                            metadata=DocMetaData(
+                                source=url,
+                                title=page["metadata"].get("title", ""),
+                            ),
+                        )
                     )
                     processed_urls.add(url)
                     new_pages += 1
@@ -300,7 +306,10 @@ class FirecrawlCrawler(BaseCrawler):
                         docs.append(
                             Document(
                                 content=result["markdown"],
-                                metadata=DocMetaData(source=url),
+                                metadata=DocMetaData(
+                                    source=url,
+                                    title=metadata.get("title", ""),
+                                ),
                             )
                         )
                 except Exception as e:
@@ -336,7 +345,7 @@ class ExaCrawler(BaseCrawler):
 
     @property
     def needs_parser(self) -> bool:
-        return False
+        return True
 
     def crawl(self, urls: List[str]) -> List[Document]:
         """Crawl the given URLs using Exa SDK.
@@ -363,12 +372,29 @@ class ExaCrawler(BaseCrawler):
         docs = []
 
         try:
-            results = exa.get_contents(urls, text=True)
-
-            for result in results.results:
-                if result.text:
-                    metadata = DocMetaData(source=result.url)
-                    docs.append(Document(content=result.text, metadata=metadata))
+            for url in urls:
+                parsed_doc_chunks = self._process_document(url)
+                if parsed_doc_chunks:
+                    docs.extend(parsed_doc_chunks)
+                    continue
+                else:
+                    results = exa.get_contents([url], livecrawl="always", text=True)
+                    result = results.results[0]
+                    if result.text:
+                        # append a NON-chunked document
+                        # (metadata.is_chunk = False, so will be chunked downstream)
+                        docs.append(
+                            Document(
+                                content=result.text,
+                                metadata=DocMetaData(
+                                    source=url,
+                                    title=getattr(result, "title", ""),
+                                    published_date=getattr(
+                                        result, "published_date", ""
+                                    ),
+                                ),
+                            )
+                        )
 
         except Exception as e:
             logging.error(f"Error retrieving content from Exa API: {e}")
@@ -399,6 +425,8 @@ class URLLoader:
             crawler_config = TrafilaturaConfig(parser=Parser(parsing_config))
 
         self.crawler = CrawlerFactory.create_crawler(crawler_config)
+        if self.crawler.needs_parser:
+            self.crawler.parser = Parser(parsing_config)
 
     def load(self) -> List[Document]:
         """Load the URLs using the specified crawler."""

{langroid-0.48.0.dist-info → langroid-0.48.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langroid
-Version: 0.48.0
+Version: 0.48.1
 Summary: Harness LLMs with Multi-Agent Programming
 Author-email: Prasad Chalasani <pchalasani@gmail.com>
 License: MIT

{langroid-0.48.0.dist-info → langroid-0.48.1.dist-info}/RECORD

@@ -1,6 +1,6 @@
 langroid/__init__.py,sha256=z_fCOLQJPOw3LLRPBlFB5-2HyCjpPgQa4m4iY5Fvb8Y,1800
 langroid/exceptions.py,sha256=OPjece_8cwg94DLPcOGA1ddzy5bGh65pxzcHMnssTz8,2995
-langroid/mytypes.py,sha256=wfb320SFnZVTv_CgcLWsvoKBXxAFfY4EISeue8MFqpQ,2912
+langroid/mytypes.py,sha256=ZW06CyhOPtemUvAGl5m4uPMHd8kEeEfwq04d4U8PntE,2975
 langroid/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/agent/__init__.py,sha256=ll0Cubd2DZ-fsCMl7e10hf9ZjFGKzphfBco396IKITY,786
 langroid/agent/base.py,sha256=U-UjdpxIFqkzRIB5-LYwHrhMSNI3sDbfnNRqIhrtsyI,79568
@@ -91,7 +91,7 @@ langroid/parsing/routing.py,sha256=-FcnlqldzL4ZoxuDwXjQPNHgBe9F9-F4R6q7b_z9CvI,1
 langroid/parsing/search.py,sha256=0NJ5-Rou_BbrHAD7O9b20bKjZJnbadjObvGm4Zq8Kis,9818
 langroid/parsing/spider.py,sha256=hAVM6wxh1pQ0EN4tI5wMBtAjIk0T-xnpi-ZUzWybhos,3258
 langroid/parsing/table_loader.py,sha256=qNM4obT_0Y4tjrxNBCNUYjKQ9oETCZ7FbolKBTcz-GM,3410
-langroid/parsing/url_loader.py,sha256=UiKlokh8AE0Qz9d4uIGPGJ-1yUfXrJjjuRBzsbxvurg,13552
+langroid/parsing/url_loader.py,sha256=Y1kFi6DoIjIxuQmMwR9SPVyHfeCJAe41eofdXUIA1fQ,14833
 langroid/parsing/urls.py,sha256=Tjzr64YsCusiYkY0LEGB5-rSuX8T2P_4DVoOFKAeKuI,8081
 langroid/parsing/utils.py,sha256=WwqzOhbQRlorbVvddDIZKv9b1KqZCBDm955lgIHDXRw,12828
 langroid/parsing/web_search.py,sha256=sARV1Tku4wiInhuCz0kRaMHcoF6Ok6CLu7vapLS8hjs,8222
@@ -127,7 +127,7 @@ langroid/vector_store/pineconedb.py,sha256=otxXZNaBKb9f_H75HTaU3lMHiaR2NUp5MqwLZ
 langroid/vector_store/postgres.py,sha256=wHPtIi2qM4fhO4pMQr95pz1ZCe7dTb2hxl4VYspGZoA,16104
 langroid/vector_store/qdrantdb.py,sha256=O6dSBoDZ0jzfeVBd7LLvsXu083xs2fxXtPa9gGX3JX4,18443
 langroid/vector_store/weaviatedb.py,sha256=Yn8pg139gOy3zkaPfoTbMXEEBCiLiYa1MU5d_3UA1K4,11847
-langroid-0.48.0.dist-info/METADATA,sha256=mhJmePv93Tsvw2Q1C2EeKrmtFzLmzxGBobBU8RAqpZs,63606
-langroid-0.48.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-langroid-0.48.0.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
-langroid-0.48.0.dist-info/RECORD,,
+langroid-0.48.1.dist-info/METADATA,sha256=5tA8WlsZ5n91APjQVDaNBVmUNwOgZ11jfdQunonoW5w,63606
+langroid-0.48.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+langroid-0.48.1.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
+langroid-0.48.1.dist-info/RECORD,,