LLM/DocsGPT
1
0
Fork 0
mirror of https://github.com/arc53/DocsGPT.git synced 2026-01-20 05:50:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: count history tokens from chunks, remove old UI setting limit (#2196)

Browse Source
This commit is contained in:
Alex
2025-12-17 01:34:17 +00:00
committed by GitHub
parent aacf281222
commit af3e16c4fc
12 changed files with 3 additions and 77 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
application/api/answer/routes/answer.py
View File

@@ -40,7 +40,6 @@ class AnswerResource(Resource, BaseAnswerResource):
"chunks": fields.Integer(
required=False, default=2, description="Number of chunks"
),
"token_limit": fields.Integer(required=False, description="Token limit"),
"retriever": fields.String(required=False, description="Retriever type"),
"api_key": fields.String(required=False, description="API key"),
"active_docs": fields.String(

1
application/api/answer/routes/stream.py
View File

@@ -40,7 +40,6 @@ class StreamResource(Resource, BaseAnswerResource):
"chunks": fields.Integer(
required=False, default=2, description="Number of chunks"
),
"token_limit": fields.Integer(required=False, description="Token limit"),
"retriever": fields.String(required=False, description="Retriever type"),
"api_key": fields.String(required=False, description="API key"),
"active_docs": fields.String(

4
application/api/answer/services/stream_processor.py
View File

@@ -420,16 +420,14 @@ class StreamProcessor:
)
def _configure_retriever(self):
history_token_limit = int(self.data.get("token_limit", 2000))
doc_token_limit = calculate_doc_token_budget(
model_id=self.model_id, history_token_limit=history_token_limit
model_id=self.model_id
)
self.retriever_config = {
"retriever_name": self.data.get("retriever", "classic"),
"chunks": int(self.data.get("chunks", 2)),
"doc_token_limit": doc_token_limit,
"history_token_limit": history_token_limit,
}
api_key = self.data.get("api_key") or self.agent_key

4
application/utils.py
View File

@@ -77,11 +77,11 @@ def count_tokens_docs(docs):
def calculate_doc_token_budget(
model_id: str = "gpt-4o", history_token_limit: int = 2000
model_id: str = "gpt-4o"
) -> int:
total_context = get_token_limit(model_id)
reserved = sum(settings.RESERVED_TOKENS.values())
doc_budget = total_context - history_token_limit - reserved
doc_budget = total_context - reserved
return max(doc_budget, 1000)