langroid 0.33.12__py3-none-any.whl → 0.33.13__py3-none-any.whl

langroid/agent/special/doc_chat_agent.py

@@ -1143,20 +1143,22 @@ class DocChatAgent(ChatAgent):
         id2_rank_semantic = {d.id(): i for i, (d, _) in enumerate(docs_and_scores)}
         id2doc = {d.id(): d for d, _ in docs_and_scores}
         # make sure we get unique docs
-        passages = [id2doc[id] for id, _ in id2_rank_semantic.items()]
+        passages = [id2doc[id] for id in id2_rank_semantic.keys()]
 
         id2_rank_bm25 = {}
         if self.config.use_bm25_search:
             # TODO: Add score threshold in config
             docs_scores = self.get_similar_chunks_bm25(query, retrieval_multiple)
+            id2doc.update({d.id(): d for d, _ in docs_scores})
             if self.config.cross_encoder_reranking_model == "":
                 # only if we're not re-ranking with a cross-encoder,
                 # we collect these ranks for Reciprocal Rank Fusion down below.
                 docs_scores = sorted(docs_scores, key=lambda x: x[1], reverse=True)
                 id2_rank_bm25 = {d.id(): i for i, (d, _) in enumerate(docs_scores)}
-                id2doc.update({d.id(): d for d, _ in docs_scores})
             else:
                 passages += [d for (d, _) in docs_scores]
+                # eliminate duplicate ids
+                passages = [id2doc[id] for id in id2doc.keys()]
 
         id2_rank_fuzzy = {}
         if self.config.use_fuzzy_match:
@@ -1174,6 +1176,8 @@ class DocChatAgent(ChatAgent):
                 id2doc.update({d.id(): d for d, _ in fuzzy_match_doc_scores})
             else:
                 passages += [d for (d, _) in fuzzy_match_doc_scores]
+                # eliminate duplicate ids
+                passages = [id2doc[id] for id in id2doc.keys()]
 
         if (
             self.config.cross_encoder_reranking_model == ""

langroid/mytypes.py

@@ -75,6 +75,16 @@ class Document(BaseModel):
     def id(self) -> str:
         return self.metadata.id
 
+    def from_string(
+        content: str,
+        source: str = "context",
+        is_chunk: bool = True,
+    ) -> "Document":
+        return Document(
+            content=content,
+            metadata=DocMetaData(source=source, is_chunk=is_chunk),
+        )
+
     def __str__(self) -> str:
         return dedent(
             f"""

{langroid-0.33.12.dist-info → langroid-0.33.13.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langroid
-Version: 0.33.12
+Version: 0.33.13
 Summary: Harness LLMs with Multi-Agent Programming
 Author-email: Prasad Chalasani <pchalasani@gmail.com>
 License: MIT

{langroid-0.33.12.dist-info → langroid-0.33.13.dist-info}/RECORD

@@ -1,6 +1,6 @@
 langroid/__init__.py,sha256=z_fCOLQJPOw3LLRPBlFB5-2HyCjpPgQa4m4iY5Fvb8Y,1800
 langroid/exceptions.py,sha256=gp6ku4ZLdXXCUQIwUNVFojJNGTzGnkevi2PLvG7HOhc,2555
-langroid/mytypes.py,sha256=ptAFxEAtiwmIfUnGisNotTe8wT9LKBf22lOfPgZoQIY,2368
+langroid/mytypes.py,sha256=vrU4eb5In8fEYwC881T4NByIaGdxiMhmY5dLxGEkyZY,2629
 langroid/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/agent/__init__.py,sha256=ll0Cubd2DZ-fsCMl7e10hf9ZjFGKzphfBco396IKITY,786
 langroid/agent/base.py,sha256=oThlrYygKDu1-bKjAfygldJ511gMKT8Z0qCrD52DdDM,77834
@@ -14,7 +14,7 @@ langroid/agent/xml_tool_message.py,sha256=6SshYZJKIfi4mkE-gIoSwjkEYekQ8GwcSiCv7a
 langroid/agent/callbacks/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/agent/callbacks/chainlit.py,sha256=RH8qUXaZE5o2WQz3WJQ1SdFtASGlxWCA6_HYz_3meDQ,20822
 langroid/agent/special/__init__.py,sha256=gik_Xtm_zV7U9s30Mn8UX3Gyuy4jTjQe9zjiE3HWmEo,1273
-langroid/agent/special/doc_chat_agent.py,sha256=zw2MvdCWRPH93d73PKh27KFiQ8sUCFPxAfLDdkxvdZQ,59301
+langroid/agent/special/doc_chat_agent.py,sha256=xXkrFb5ty3NoKfgC0CZqolxmUP1j_sXqywQ_VNxOrOE,59505
 langroid/agent/special/lance_doc_chat_agent.py,sha256=s8xoRs0gGaFtDYFUSIRchsgDVbS5Q3C2b2mr3V1Fd-Q,10419
 langroid/agent/special/lance_tools.py,sha256=qS8x4wi8mrqfbYV2ztFzrcxyhHQ0ZWOc-zkYiH7awj0,2105
 langroid/agent/special/relevance_extractor_agent.py,sha256=zIx8GUdVo1aGW6ASla0NPQjYYIpmriK_TYMijqAx3F8,4796
@@ -121,7 +121,7 @@ langroid/vector_store/lancedb.py,sha256=b3_vWkTjG8mweZ7ZNlUD-NjmQP_rLBZfyKWcxt2v
 langroid/vector_store/meilisearch.py,sha256=6frB7GFWeWmeKzRfLZIvzRjllniZ1cYj3HmhHQICXLs,11663
 langroid/vector_store/momento.py,sha256=UNHGT6jXuQtqY9f6MdqGU14bVnS0zHgIJUa30ULpUJo,10474
 langroid/vector_store/qdrantdb.py,sha256=HRLCt-FG8y4718omwpFaQZnWeYxPj0XCwS4tjokI1sU,18116
-langroid-0.33.12.dist-info/METADATA,sha256=tUh6elP7kcHfPS2dUXBE-gZ6vLoBfFzYsmg7nr2oCrg,59016
-langroid-0.33.12.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-langroid-0.33.12.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
-langroid-0.33.12.dist-info/RECORD,,
+langroid-0.33.13.dist-info/METADATA,sha256=6D80UFs_rmlz89V6OEStdRBVECC-B8iLtBLTghgRg5w,59016
+langroid-0.33.13.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+langroid-0.33.13.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
+langroid-0.33.13.dist-info/RECORD,,