oobabooga/text-generation-webui
1
0
Fork 0
mirror of https://github.com/oobabooga/text-generation-webui.git synced 2026-03-18 03:14:39 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity

UI: Fix tool calling for GPT-OSS and Continue

Browse source
This commit is contained in:
oobabooga 2026-03-12 22:17:20 -03:00
parent 1b7e6c5705
commit a09f21b9de
3 changed files with 64 additions and 32 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
extensions/openai/utils.py
View file

@ -120,12 +120,14 @@ def _parseChannelToolCalls(answer: str, tool_names: list[str]):
"""Parse channel-based tool calls used by GPT-OSS and similar models.
Format:
<|start|>assistant to=functions.func_name<|channel|>commentary json<|message|>{"arg": "value"}
or:
<|channel|>commentary to=functions.func_name <|constrain|>json<|message|>{"arg": "value"}
"""
matches = []
start_pos = None
for m in re.finditer(
r'<\|channel\|>commentary to=functions\.([^<\s]+)\s*(?:<\|constrain\|>json)?<\|message\|>',
r'<\|channel\|>\w+ to=functions\.([^<\s]+).*?<\|message\|>',
answer
):
func_name = m.group(1).strip()
@ -137,7 +139,8 @@ def _parseChannelToolCalls(answer: str, tool_names: list[str]):
try:
arguments = json.loads(json_str)
if start_pos is None:
start_pos = m.start()
prefix = answer.rfind('<|start|>assistant', 0, m.start())
start_pos = prefix if prefix != -1 else m.start()
matches.append({
"type": "function",
"function": {

88
modules/chat.py
View file

@ -81,6 +81,13 @@ jinja_env = ImmutableSandboxedEnvironment(
)
jinja_env.globals["strftime_now"] = strftime_now
def _raise_exception(message):
raise ValueError(message)
jinja_env.globals["raise_exception"] = _raise_exception
_template_cache = {}
@ -1048,16 +1055,23 @@ def chatbot_wrapper(text, state, regenerate=False, _continue=False, loading_mess
yield output
if _continue:
# Reprocess the entire internal text for extensions (like translation)
full_internal = output['internal'][-1][1]
if state['mode'] in ['chat', 'chat-instruct']:
full_visible = re.sub("(<USER>|<user>|{{user}})", state['name1'], full_internal)
else:
full_visible = full_internal
# Reprocess the entire internal text for extensions (like translation).
# Skip the rebuild when the visible text contains <tool_call> markers,
# since those only exist in visible (internal is cleared after each tool
# execution) and rebuilding from internal would destroy them.
if '<tool_call>' not in output['visible'][-1][1]:
full_internal = output['internal'][-1][1]
if state['mode'] in ['chat', 'chat-instruct']:
full_visible = re.sub("(<USER>|<user>|{{user}})", state['name1'], full_internal)
else:
full_visible = full_internal
full_visible = html.escape(full_visible)
if not state.get('_skip_output_extensions'):
output['visible'][-1][1] = apply_extensions('output', full_visible, state, is_chat=True)
full_visible = html.escape(full_visible)
if not state.get('_skip_output_extensions'):
output['visible'][-1][1] = apply_extensions('output', full_visible, state, is_chat=True)
else:
if not state.get('_skip_output_extensions'):
output['visible'][-1][1] = apply_extensions('output', output['visible'][-1][1], state, is_chat=True)
else:
if not state.get('_skip_output_extensions'):
output['visible'][-1][1] = apply_extensions('output', output['visible'][-1][1], state, is_chat=True)
@ -1222,6 +1236,18 @@ def generate_chat_reply_wrapper(text, state, regenerate=False, _continue=False):
# so we can extract thinking content from just this turn's output.
_model_visible = history['visible'][-1][1]
# Recover visible_prefix from existing visible text (e.g. on Continue
# after a previous session had tool calls). Extract all <tool_call>
# blocks and any text between them (thinking blocks, intermediate text).
if not visible_prefix and _model_visible:
tc_matches = list(re.finditer(r'<tool_call>.*?</tool_call>', _model_visible, re.DOTALL))
if tc_matches:
prefix_end = tc_matches[-1].end()
prefix = _model_visible[:prefix_end].strip()
if prefix:
visible_prefix = [prefix]
_model_visible = _model_visible[prefix_end:].strip()
# Re-apply visible prefix to the final state after streaming completes.
# This is safe because we're no longer sharing the object with chatbot_wrapper.
if visible_prefix:
@ -1244,20 +1270,35 @@ def generate_chat_reply_wrapper(text, state, regenerate=False, _continue=False):
meta = history.get('metadata', {})
seq = meta.setdefault(f'assistant_{row_idx}', {}).setdefault('tool_sequence', [])
# Serialize tool calls
def _render():
return chat_html_wrapper(history, state['name1'], state['name2'], state['mode'], state['chat_style'], state['character_menu'])
# Serialize tool calls and build display headers in one pass
serialized = []
tc_headers = []
for tc in parsed_calls:
tc['id'] = generate_tool_call_id()
args = tc['function'].get('arguments', {})
fn_name = tc['function']['name']
fn_args = tc['function'].get('arguments', {})
serialized.append({
'id': tc['id'],
'type': 'function',
'function': {
'name': tc['function']['name'],
'arguments': json.dumps(args) if isinstance(args, dict) else args
'name': fn_name,
'arguments': json.dumps(fn_args) if isinstance(fn_args, dict) else fn_args
}
})
if isinstance(fn_args, dict) and fn_args:
args_summary = ', '.join(f'{k}={json.dumps(v, ensure_ascii=False)}' for k, v in fn_args.items())
elif isinstance(fn_args, dict):
args_summary = ''
else:
args_summary = str(fn_args)
tc_headers.append(f'{fn_name}({args_summary})')
seq.append({'tool_calls': serialized})
# Clear internal (raw tool markup)
@ -1274,24 +1315,11 @@ def generate_chat_reply_wrapper(text, state, regenerate=False, _continue=False):
if intermediate and intermediate.strip():
visible_prefix.append(intermediate.strip())
# Build args summaries and show placeholder accordions with "..."
# before execution starts (tool calls may be slow, e.g. web search).
tc_headers = []
for tc in parsed_calls:
fn_name = tc['function']['name']
fn_args = tc['function'].get('arguments', {})
if isinstance(fn_args, dict) and fn_args:
args_summary = ', '.join(f'{k}={json.dumps(v, ensure_ascii=False)}' for k, v in fn_args.items())
elif isinstance(fn_args, dict):
args_summary = ''
else:
args_summary = str(fn_args)
tc_headers.append(f'{fn_name}({args_summary})')
# Show placeholder accordions with "..." before execution starts
# (tool calls may be slow, e.g. web search).
pending_placeholders = [f'<tool_call>{h}\n...\n</tool_call>' for h in tc_headers]
history['visible'][-1][1] = '\n\n'.join(visible_prefix + pending_placeholders)
yield chat_html_wrapper(history, state['name1'], state['name2'], state['mode'], state['chat_style'], state['character_menu']), history
yield _render(), history
# Execute tools, store results, and replace placeholders with real results
for i, tc in enumerate(parsed_calls):
@ -1308,7 +1336,7 @@ def generate_chat_reply_wrapper(text, state, regenerate=False, _continue=False):
# Replace the placeholder with the real result
pending_placeholders[i] = f'<tool_call>{tc_headers[i]}\n{pretty_result}\n</tool_call>'
history['visible'][-1][1] = '\n\n'.join(visible_prefix + pending_placeholders)
yield chat_html_wrapper(history, state['name1'], state['name2'], state['mode'], state['chat_style'], state['character_menu']), history
yield _render(), history
# Move completed tool calls into visible_prefix for next turns
visible_prefix.extend(pending_placeholders)

1
modules/reasoning.py
View file

@ -5,6 +5,7 @@ import html as html_module
THINKING_FORMATS = [
('<think>', '</think>', None),
('<|channel|>analysis<|message|>', '<|end|>', '<|start|>assistant<|channel|>final<|message|>'),
('<|channel|>commentary<|message|>', '<|end|>', '<|start|>assistant<|channel|>final<|message|>'),
('<seed:think>', '</seed:think>', None),
('<|think|>', '<|end|>', '<|content|>'), # Solar Open
('Thinking Process:', '</think>', None), # Qwen3.5 verbose thinking outside tags