[Chat] Add prompt_truncation param

CHANGELOG.md

@@ -1,34 +1,45 @@
 # Changelog
 
+## 4.26
+
+- [#296] (https://github.com/cohere-ai/cohere-python/pull/301)
+  - Chat: add support for prompt_truncation param
+
 ## 4.25 
 - [#303] (https://github.com/cohere-ai/cohere-python/pull/303)
   - Allow uploading of evaluation data
 
 ## 4.24
+
 - [#296] (https://github.com/cohere-ai/cohere-python/pull/296)
   - Allow passing of delimiter for csv
-  
+
 ## 4.23
+
 - [#294] (https://github.com/cohere-ai/cohere-python/pull/294)
   - Allow passing of ParseInfo for datasets
 
 ## 4.22
+
 - [#292] (https://github.com/cohere-ai/cohere-python/pull/292)
   - Add search query only parameter
-  - Add documents to generate citations 
-  - Add connectors to generate citations 
+  - Add documents to generate citations
+  - Add connectors to generate citations
   - Add citation quality
 
 ## 4.21
+
 - [#287] (https://github.com/cohere-ai/cohere-python/pull/287)
   - Remove deprecated chat "query" parameter including inside chat_history parameter
   - Support event-type for chat streaming
 
 ## 4.20.2
+
 - [#284] (https://github.com/cohere-ai/cohere-python/pull/284)
   - Rename dataset urls to download_urls
 
 ## 4.20.1
+
 - [#279] (https://github.com/cohere-ai/cohere-python/pull/279)
   - Fix dataset listing key error
 
@@ -64,7 +75,7 @@
 - [#262](https://github.com/cohere-ai/cohere-python/pull/262)
   - Deprecate embed custom models
 
-## 4.15 
+## 4.15
 
 - [#256](https://github.com/cohere-ai/cohere-python/pull/256)
   - Add rerank finetuning endpoint

cohere/client.py

@@ -244,6 +244,7 @@ def chat(
         search_queries_only: Optional[bool] = None,
         documents: Optional[List[Dict[str, Any]]] = None,
         citation_quality: Optional[str] = None,
+        prompt_truncation: Optional[str] = None,
         connectors: Optional[List[Dict[str, Any]]] = None,
     ) -> Union[Chat, StreamingChat]:
         """Returns a Chat object with the query reply.
@@ -288,6 +289,7 @@ def chat(
                 ],
             connectors (List[Dict[str, str]]): (Optional) When specified, the model's reply will be enriched with information found by quering each of the connectors (RAG). Example: connectors=[{"id": "web-search"}]
             citation_quality (str): (Optional) Dictates the approach taken to generating citations by allowing the user to specify whether they want "accurate" results or "fast" results. Defaults to "accurate".
+            prompt_truncation (str): (Optional) Dictates how the prompt will be constructed. With `prompt_truncation` set to "AUTO", some elements from `chat_history` and `documents` will be dropped in attempt to construct a prompt that fits within the model's context length limit. With `prompt_truncation` set to "OFF", no elements will be dropped. If the sum of the inputs exceeds the model's context length limit, a `TooManyTokens` error will be returned.
         Returns:
             a Chat object if stream=False, or a StreamingChat object if stream=True
 
@@ -376,6 +378,9 @@ def chat(
         }
         if citation_quality is not None:
             json_body["citation_quality"] = citation_quality
+        if prompt_truncation is not None:
+            json_body["prompt_truncation"] = prompt_truncation
+
         response = self._request(cohere.CHAT_URL, json=json_body, stream=stream)
 
         if stream:

cohere/client_async.py

@@ -226,6 +226,7 @@ async def chat(
         search_queries_only: Optional[bool] = None,
         documents: Optional[List[Dict[str, Any]]] = None,
         citation_quality: Optional[str] = None,
+        prompt_truncation: Optional[str] = None,
         connectors: Optional[List[Dict[str, Any]]] = None,
     ) -> Union[AsyncChat, StreamingChat]:
         if message is None:
@@ -253,6 +254,8 @@ async def chat(
         }
         if citation_quality is not None:
             json_body["citation_quality"] = citation_quality
+        if prompt_truncation is not None:
+            json_body["prompt_truncation"] = prompt_truncation
 
         response = await self._request(cohere.CHAT_URL, json=json_body, stream=stream)
 

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "cohere"
-version = "4.25"
+version = "4.26"
 description = ""
 authors = ["Cohere"]
 readme = "README.md"