162 lines
7.6 KiB
Python
162 lines
7.6 KiB
Python
#!/usr/bin/env python3
|
|
"""Make Open WebUI finish with an answer when its internal tool budget is used.
|
|
|
|
The pinned Open WebUI revision otherwise ends the request with only
|
|
``Tool-call limit reached``. The patch is deliberately assertion-based: an
|
|
upstream source change makes the image build fail instead of silently applying
|
|
the modification at the wrong location.
|
|
"""
|
|
|
|
from pathlib import Path
|
|
import sys
|
|
|
|
|
|
if len(sys.argv) != 2:
|
|
raise SystemExit("usage: patch_tool_finalization.py MIDDLEWARE_PY")
|
|
|
|
path = Path(sys.argv[1])
|
|
source = path.read_text()
|
|
|
|
patch_marker = "The tool budget for this turn is exhausted. Do not call"
|
|
if patch_marker in source:
|
|
raise SystemExit("Open WebUI tool-finalization patch is already present")
|
|
|
|
needle = """ tool_call_iterations = 0
|
|
max_tool_call_iterations = getattr(
|
|
"""
|
|
replacement = """ tool_call_iterations = 0
|
|
# Open WebUI counts batches, while one model turn may request many
|
|
# functions in parallel. Bound actual executions as well so a small
|
|
# local model cannot expand eight rounds into dozens of API calls.
|
|
tool_call_executions = 0
|
|
max_tool_call_executions = 12
|
|
max_tool_call_iterations = getattr(
|
|
"""
|
|
if source.count(needle) != 1:
|
|
raise SystemExit(
|
|
f"expected exactly one Open WebUI tool counter anchor, found {source.count(needle)}"
|
|
)
|
|
source = source.replace(needle, replacement)
|
|
|
|
needle = """ response_tool_calls = tool_calls.pop(0)
|
|
|
|
# Append function_call items for each tool call
|
|
"""
|
|
replacement = """ response_tool_calls = tool_calls.pop(0)
|
|
|
|
remaining_tool_calls = max(
|
|
0, max_tool_call_executions - tool_call_executions
|
|
)
|
|
skipped_tool_calls = response_tool_calls[remaining_tool_calls:]
|
|
response_tool_calls = response_tool_calls[:remaining_tool_calls]
|
|
if skipped_tool_calls:
|
|
skipped_ids = {call.get('id', '') for call in skipped_tool_calls}
|
|
# Responses API streaming may already have exposed all calls
|
|
# in `output`. Remove deliberately skipped calls so the next
|
|
# completion never receives an orphan function call.
|
|
output[:] = [
|
|
item
|
|
for item in output
|
|
if not (
|
|
item.get('type') == 'function_call'
|
|
and item.get('call_id', '') in skipped_ids
|
|
)
|
|
]
|
|
tool_call_executions += len(response_tool_calls)
|
|
|
|
# Append function_call items for each tool call
|
|
"""
|
|
if source.count(needle) != 1:
|
|
raise SystemExit(
|
|
f"expected exactly one Open WebUI tool batch anchor, found {source.count(needle)}"
|
|
)
|
|
source = source.replace(needle, replacement)
|
|
|
|
needle = """ res = await generate_chat_completion(
|
|
request,
|
|
new_form_data,
|
|
user,
|
|
bypass_system_prompt=True,
|
|
)
|
|
"""
|
|
|
|
replacement = """ # The upstream loop otherwise stops with only
|
|
# `Tool-call limit reached` after executing this batch. Reserve
|
|
# the final completion for synthesis: keep every accumulated
|
|
# result, but expose no tools to the model and explicitly require
|
|
# a useful, evidence-bounded answer.
|
|
force_final_response = (
|
|
(
|
|
max_tool_call_iterations is not None
|
|
and tool_call_iterations >= max_tool_call_iterations
|
|
)
|
|
or tool_call_executions >= max_tool_call_executions
|
|
)
|
|
if force_final_response:
|
|
new_form_data.pop('tools', None)
|
|
new_form_data.pop('tool_ids', None)
|
|
new_form_data.pop('tool_choice', None)
|
|
final_metadata = dict(metadata)
|
|
final_metadata['tools'] = {}
|
|
final_metadata['tool_ids'] = []
|
|
final_metadata['tool_servers'] = []
|
|
new_form_data['metadata'] = final_metadata
|
|
new_form_data['messages'] = add_or_update_system_message(
|
|
'The tool budget for this turn is exhausted. Do not call '
|
|
'or imitate any more tools. Give the user a concise final '
|
|
'answer now using only the tool results already present. '
|
|
'Explicitly distinguish verified findings from inference, '
|
|
'and state what could not be verified. Never end without a '
|
|
'visible answer.',
|
|
new_form_data['messages'],
|
|
append=True,
|
|
)
|
|
# A final user-role instruction is deliberately stronger
|
|
# than another system suffix after a long tool-call pattern.
|
|
# It is request-local and is not added to the saved chat.
|
|
new_form_data['messages'].append(
|
|
{
|
|
'role': 'user',
|
|
'content': (
|
|
'The research phase is finished. Answer my original '
|
|
'question now in normal prose. Do not output XML, JSON, '
|
|
'function names, tool_call blocks, or requests for more '
|
|
'files. Use the available evidence, mention uncertainty, '
|
|
'and provide a useful final conclusion.'
|
|
),
|
|
}
|
|
)
|
|
|
|
res = await generate_chat_completion(
|
|
request,
|
|
new_form_data,
|
|
user,
|
|
bypass_system_prompt=True,
|
|
)
|
|
"""
|
|
|
|
if source.count(needle) != 1:
|
|
raise SystemExit(
|
|
f"expected exactly one Open WebUI continuation anchor, found {source.count(needle)}"
|
|
)
|
|
source = source.replace(needle, replacement)
|
|
|
|
needle = """ await stream_body_handler(res, new_form_data)
|
|
output[:0] = prior_output
|
|
"""
|
|
replacement = """ await stream_body_handler(res, new_form_data)
|
|
if force_final_response:
|
|
# A model may still print tool-shaped output after the
|
|
# schemas were removed. It must not trigger the upstream
|
|
# hard-error path or another execution attempt.
|
|
tool_calls.clear()
|
|
output[:0] = prior_output
|
|
"""
|
|
if source.count(needle) != 1:
|
|
raise SystemExit(
|
|
f"expected exactly one Open WebUI stream anchor, found {source.count(needle)}"
|
|
)
|
|
source = source.replace(needle, replacement)
|
|
|
|
path.write_text(source)
|