mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-04-25 00:51:20 +00:00
feat: call_llm/async_call_llm + config slots + migrate all consumers
Add centralized call_llm() and async_call_llm() functions that own the full LLM request lifecycle: 1. Resolve provider + model from task config or explicit args 2. Get or create a cached client for that provider 3. Format request args (max_tokens handling, provider extra_body) 4. Make the API call with max_tokens/max_completion_tokens retry 5. Return the response Config: expanded auxiliary section with provider:model slots for all tasks (compression, vision, web_extract, session_search, skills_hub, mcp, flush_memories). Config version bumped to 7. Migrated all auxiliary consumers: - context_compressor.py: uses call_llm(task='compression') - vision_tools.py: uses async_call_llm(task='vision') - web_tools.py: uses async_call_llm(task='web_extract') - session_search_tool.py: uses async_call_llm(task='session_search') - browser_tool.py: uses call_llm(task='vision'/'web_extract') - mcp_tool.py: uses call_llm(task='mcp') - skills_guard.py: uses call_llm(provider='openrouter') - run_agent.py flush_memories: uses call_llm(task='flush_memories') Tests updated for context_compressor and MCP tool. Some test mocks still need updating (15 remaining failures from mock pattern changes, 2 pre-existing).
This commit is contained in:
teknium1
parent
013cc4d2fc
commit
0aa31cd3cb
13 changed files with 552 additions and 375 deletions
|
|
@ -47,8 +47,7 @@ import re
|
|||
import asyncio
|
||||
from typing import List, Dict, Any, Optional
|
||||
from firecrawl import Firecrawl
|
||||
from openai import AsyncOpenAI
|
||||
from agent.auxiliary_client import get_async_text_auxiliary_client
|
||||
from agent.auxiliary_client import async_call_llm
|
||||
from tools.debug_helpers import DebugSession
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
|
@ -83,15 +82,8 @@ def _get_firecrawl_client():
|
|||
|
||||
DEFAULT_MIN_LENGTH_FOR_SUMMARIZATION = 5000
|
||||
|
||||
# Resolve async auxiliary client at module level.
|
||||
# Handles Codex Responses API adapter transparently.
|
||||
_aux_async_client, _DEFAULT_SUMMARIZER_MODEL = get_async_text_auxiliary_client("web_extract")
|
||||
|
||||
# Allow per-task override via config.yaml auxiliary.web_extract_model
|
||||
DEFAULT_SUMMARIZER_MODEL = (
|
||||
os.getenv("AUXILIARY_WEB_EXTRACT_MODEL", "").strip()
|
||||
or _DEFAULT_SUMMARIZER_MODEL
|
||||
)
|
||||
# Allow per-task override via env var
|
||||
DEFAULT_SUMMARIZER_MODEL = os.getenv("AUXILIARY_WEB_EXTRACT_MODEL", "").strip() or None
|
||||
|
||||
_debug = DebugSession("web_tools", env_var="WEB_TOOLS_DEBUG")
|
||||
|
||||
|
|
@ -249,22 +241,22 @@ Create a markdown summary that captures all key information in a well-organized,
|
|||
|
||||
for attempt in range(max_retries):
|
||||
try:
|
||||
if _aux_async_client is None:
|
||||
logger.warning("No auxiliary model available for web content processing")
|
||||
return None
|
||||
from agent.auxiliary_client import get_auxiliary_extra_body, auxiliary_max_tokens_param
|
||||
_extra = get_auxiliary_extra_body()
|
||||
response = await _aux_async_client.chat.completions.create(
|
||||
model=model,
|
||||
messages=[
|
||||
call_kwargs = {
|
||||
"task": "web_extract",
|
||||
"messages": [
|
||||
{"role": "system", "content": system_prompt},
|
||||
{"role": "user", "content": user_prompt}
|
||||
],
|
||||
temperature=0.1,
|
||||
**auxiliary_max_tokens_param(max_tokens),
|
||||
**({} if not _extra else {"extra_body": _extra}),
|
||||
)
|
||||
"temperature": 0.1,
|
||||
"max_tokens": max_tokens,
|
||||
}
|
||||
if model:
|
||||
call_kwargs["model"] = model
|
||||
response = await async_call_llm(**call_kwargs)
|
||||
return response.choices[0].message.content.strip()
|
||||
except RuntimeError:
|
||||
logger.warning("No auxiliary model available for web content processing")
|
||||
return None
|
||||
except Exception as api_error:
|
||||
last_error = api_error
|
||||
if attempt < max_retries - 1:
|
||||
|
|
@ -368,25 +360,18 @@ Synthesize these into ONE cohesive, comprehensive summary that:
|
|||
Create a single, unified markdown summary."""
|
||||
|
||||
try:
|
||||
if _aux_async_client is None:
|
||||
logger.warning("No auxiliary model for synthesis, concatenating summaries")
|
||||
fallback = "\n\n".join(summaries)
|
||||
if len(fallback) > max_output_size:
|
||||
fallback = fallback[:max_output_size] + "\n\n[... truncated ...]"
|
||||
return fallback
|
||||
|
||||
from agent.auxiliary_client import get_auxiliary_extra_body, auxiliary_max_tokens_param
|
||||
_extra = get_auxiliary_extra_body()
|
||||
response = await _aux_async_client.chat.completions.create(
|
||||
model=model,
|
||||
messages=[
|
||||
call_kwargs = {
|
||||
"task": "web_extract",
|
||||
"messages": [
|
||||
{"role": "system", "content": "You synthesize multiple summaries into one cohesive, comprehensive summary. Be thorough but concise."},
|
||||
{"role": "user", "content": synthesis_prompt}
|
||||
],
|
||||
temperature=0.1,
|
||||
**auxiliary_max_tokens_param(20000),
|
||||
**({} if not _extra else {"extra_body": _extra}),
|
||||
)
|
||||
"temperature": 0.1,
|
||||
"max_tokens": 20000,
|
||||
}
|
||||
if model:
|
||||
call_kwargs["model"] = model
|
||||
response = await async_call_llm(**call_kwargs)
|
||||
final_summary = response.choices[0].message.content.strip()
|
||||
|
||||
# Enforce hard cap
|
||||
|
|
@ -713,8 +698,8 @@ async def web_extract_tool(
|
|||
debug_call_data["pages_extracted"] = pages_extracted
|
||||
debug_call_data["original_response_size"] = len(json.dumps(response))
|
||||
|
||||
# Process each result with LLM if enabled and auxiliary client is available
|
||||
if use_llm_processing and _aux_async_client is not None:
|
||||
# Process each result with LLM if enabled
|
||||
if use_llm_processing:
|
||||
logger.info("Processing extracted content with LLM (parallel)...")
|
||||
debug_call_data["processing_applied"].append("llm_processing")
|
||||
|
||||
|
|
@ -780,10 +765,6 @@ async def web_extract_tool(
|
|||
else:
|
||||
logger.warning("%s (no content to process)", url)
|
||||
else:
|
||||
if use_llm_processing and _aux_async_client is None:
|
||||
logger.warning("LLM processing requested but no auxiliary model available, returning raw content")
|
||||
debug_call_data["processing_applied"].append("llm_processing_unavailable")
|
||||
|
||||
# Print summary of extracted pages for debugging (original behavior)
|
||||
for result in response.get('results', []):
|
||||
url = result.get('url', 'Unknown URL')
|
||||
|
|
@ -1013,8 +994,8 @@ async def web_crawl_tool(
|
|||
debug_call_data["pages_crawled"] = pages_crawled
|
||||
debug_call_data["original_response_size"] = len(json.dumps(response))
|
||||
|
||||
# Process each result with LLM if enabled and auxiliary client is available
|
||||
if use_llm_processing and _aux_async_client is not None:
|
||||
# Process each result with LLM if enabled
|
||||
if use_llm_processing:
|
||||
logger.info("Processing crawled content with LLM (parallel)...")
|
||||
debug_call_data["processing_applied"].append("llm_processing")
|
||||
|
||||
|
|
@ -1080,10 +1061,6 @@ async def web_crawl_tool(
|
|||
else:
|
||||
logger.warning("%s (no content to process)", page_url)
|
||||
else:
|
||||
if use_llm_processing and _aux_async_client is None:
|
||||
logger.warning("LLM processing requested but no auxiliary model available, returning raw content")
|
||||
debug_call_data["processing_applied"].append("llm_processing_unavailable")
|
||||
|
||||
# Print summary of crawled pages for debugging (original behavior)
|
||||
for result in response.get('results', []):
|
||||
page_url = result.get('url', 'Unknown URL')
|
||||
|
|
@ -1138,7 +1115,15 @@ def check_firecrawl_api_key() -> bool:
|
|||
|
||||
def check_auxiliary_model() -> bool:
|
||||
"""Check if an auxiliary text model is available for LLM content processing."""
|
||||
return _aux_async_client is not None
|
||||
try:
|
||||
from agent.auxiliary_client import resolve_provider_client
|
||||
for p in ("openrouter", "nous", "custom", "codex"):
|
||||
client, _ = resolve_provider_client(p)
|
||||
if client is not None:
|
||||
return True
|
||||
return False
|
||||
except Exception:
|
||||
return False
|
||||
|
||||
|
||||
def get_debug_session_info() -> Dict[str, Any]:
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue