Files
aiagent/backend/app/core/tools_bootstrap.py
renjianbo 1f7c136544 feat: #33 多模态Agent — 图片识别/视觉理解/语音转文字/文字转语音
后端新增 4 个内置工具: image_ocr (Tesseract OCR)、image_vision (GPT-4o 多模态视觉)、
speech_to_text (Whisper API)、text_to_speech (TTS API)。
前端 AgentChatPreview 增加录音上传和语音朗读交互。

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-05-06 22:02:19 +08:00

172 lines
7.6 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""确保内置工具注册到 tool_registryAPI 进程与 Celery Worker 均会导入执行流)。"""
from __future__ import annotations
import logging
import os
logger = logging.getLogger(__name__)
_registered = False
_EXPECTED_BUILTIN = 43
def ensure_builtin_tools_registered() -> None:
"""幂等:注册所有内置工具,供工作流 LLM 节点使用。"""
global _registered
if _registered:
return
from app.services.tool_registry import tool_registry
from app.services.builtin_tools import (
http_request_tool,
file_read_tool,
file_write_tool,
text_analyze_tool,
datetime_tool,
math_calculate_tool,
system_info_tool,
json_process_tool,
database_query_tool,
adb_log_tool,
schedule_create_tool,
schedule_list_tool,
schedule_delete_tool,
crypto_util_tool,
random_generate_tool,
send_email_tool,
url_parse_tool,
regex_test_tool,
agent_call_tool,
code_execute_tool,
git_operation_tool,
web_search_tool,
pdf_generate_tool,
project_scaffold_tool,
task_plan_tool,
excel_process_tool,
browser_use_tool,
docker_manage_tool,
deploy_push_tool,
agent_create_tool,
tool_register_tool,
capability_check_tool,
code_tool_create_tool,
extension_log_tool,
self_review_tool,
knowledge_graph_search_tool,
knowledge_graph_add_tool,
entity_search_tool,
learning_path_tool,
image_ocr_tool,
image_vision_tool,
speech_to_text_tool,
text_to_speech_tool,
HTTP_REQUEST_SCHEMA,
FILE_READ_SCHEMA,
FILE_WRITE_SCHEMA,
TEXT_ANALYZE_SCHEMA,
DATETIME_SCHEMA,
MATH_CALCULATE_SCHEMA,
SYSTEM_INFO_SCHEMA,
JSON_PROCESS_SCHEMA,
DATABASE_QUERY_SCHEMA,
ADB_LOG_SCHEMA,
SCHEDULE_CREATE_SCHEMA,
SCHEDULE_LIST_SCHEMA,
SCHEDULE_DELETE_SCHEMA,
CRYPTO_UTIL_SCHEMA,
RANDOM_GENERATE_SCHEMA,
SEND_EMAIL_SCHEMA,
URL_PARSE_SCHEMA,
REGEX_TEST_SCHEMA,
AGENT_CALL_SCHEMA,
CODE_EXECUTE_SCHEMA,
GIT_OPERATION_SCHEMA,
WEB_SEARCH_SCHEMA,
PDF_GENERATE_SCHEMA,
PROJECT_SCAFFOLD_SCHEMA,
TASK_PLAN_SCHEMA,
EXCEL_PROCESS_SCHEMA,
BROWSER_USE_SCHEMA,
DOCKER_MANAGE_SCHEMA,
DEPLOY_PUSH_SCHEMA,
AGENT_CREATE_SCHEMA,
TOOL_REGISTER_SCHEMA,
CAPABILITY_CHECK_SCHEMA,
CODE_TOOL_CREATE_SCHEMA,
EXTENSION_LOG_SCHEMA,
SELF_REVIEW_SCHEMA,
KNOWLEDGE_GRAPH_SEARCH_SCHEMA,
KNOWLEDGE_GRAPH_ADD_SCHEMA,
ENTITY_SEARCH_SCHEMA,
LEARNING_PATH_SCHEMA,
IMAGE_OCR_SCHEMA,
IMAGE_VISION_SCHEMA,
SPEECH_TO_TEXT_SCHEMA,
TEXT_TO_SPEECH_SCHEMA,
)
tool_registry.register_builtin_tool("http_request", http_request_tool, HTTP_REQUEST_SCHEMA)
tool_registry.register_builtin_tool("file_read", file_read_tool, FILE_READ_SCHEMA)
tool_registry.register_builtin_tool("file_write", file_write_tool, FILE_WRITE_SCHEMA)
tool_registry.register_builtin_tool("text_analyze", text_analyze_tool, TEXT_ANALYZE_SCHEMA)
tool_registry.register_builtin_tool("datetime", datetime_tool, DATETIME_SCHEMA)
tool_registry.register_builtin_tool("math_calculate", math_calculate_tool, MATH_CALCULATE_SCHEMA)
tool_registry.register_builtin_tool("system_info", system_info_tool, SYSTEM_INFO_SCHEMA)
tool_registry.register_builtin_tool("json_process", json_process_tool, JSON_PROCESS_SCHEMA)
tool_registry.register_builtin_tool("database_query", database_query_tool, DATABASE_QUERY_SCHEMA)
tool_registry.register_builtin_tool("adb_log", adb_log_tool, ADB_LOG_SCHEMA)
tool_registry.register_builtin_tool("schedule_create", schedule_create_tool, SCHEDULE_CREATE_SCHEMA)
tool_registry.register_builtin_tool("schedule_list", schedule_list_tool, SCHEDULE_LIST_SCHEMA)
tool_registry.register_builtin_tool("schedule_delete", schedule_delete_tool, SCHEDULE_DELETE_SCHEMA)
tool_registry.register_builtin_tool("crypto_util", crypto_util_tool, CRYPTO_UTIL_SCHEMA)
tool_registry.register_builtin_tool("random_generate", random_generate_tool, RANDOM_GENERATE_SCHEMA)
tool_registry.register_builtin_tool("send_email", send_email_tool, SEND_EMAIL_SCHEMA)
tool_registry.register_builtin_tool("url_parse", url_parse_tool, URL_PARSE_SCHEMA)
tool_registry.register_builtin_tool("regex_test", regex_test_tool, REGEX_TEST_SCHEMA)
tool_registry.register_builtin_tool("agent_call", agent_call_tool, AGENT_CALL_SCHEMA)
tool_registry.register_builtin_tool("code_execute", code_execute_tool, CODE_EXECUTE_SCHEMA)
tool_registry.register_builtin_tool("git_operation", git_operation_tool, GIT_OPERATION_SCHEMA)
tool_registry.register_builtin_tool("web_search", web_search_tool, WEB_SEARCH_SCHEMA)
tool_registry.register_builtin_tool("pdf_generate", pdf_generate_tool, PDF_GENERATE_SCHEMA)
tool_registry.register_builtin_tool("project_scaffold", project_scaffold_tool, PROJECT_SCAFFOLD_SCHEMA)
tool_registry.register_builtin_tool("task_plan", task_plan_tool, TASK_PLAN_SCHEMA)
tool_registry.register_builtin_tool("excel_process", excel_process_tool, EXCEL_PROCESS_SCHEMA)
tool_registry.register_builtin_tool("browser_use", browser_use_tool, BROWSER_USE_SCHEMA)
tool_registry.register_builtin_tool("docker_manage", docker_manage_tool, DOCKER_MANAGE_SCHEMA)
tool_registry.register_builtin_tool("deploy_push", deploy_push_tool, DEPLOY_PUSH_SCHEMA)
tool_registry.register_builtin_tool("agent_create", agent_create_tool, AGENT_CREATE_SCHEMA)
tool_registry.register_builtin_tool("tool_register", tool_register_tool, TOOL_REGISTER_SCHEMA)
tool_registry.register_builtin_tool("capability_check", capability_check_tool, CAPABILITY_CHECK_SCHEMA)
tool_registry.register_builtin_tool("code_tool_create", code_tool_create_tool, CODE_TOOL_CREATE_SCHEMA)
tool_registry.register_builtin_tool("extension_log", extension_log_tool, EXTENSION_LOG_SCHEMA)
tool_registry.register_builtin_tool("self_review", self_review_tool, SELF_REVIEW_SCHEMA)
tool_registry.register_builtin_tool("knowledge_graph_search", knowledge_graph_search_tool, KNOWLEDGE_GRAPH_SEARCH_SCHEMA)
tool_registry.register_builtin_tool("knowledge_graph_add", knowledge_graph_add_tool, KNOWLEDGE_GRAPH_ADD_SCHEMA)
tool_registry.register_builtin_tool("entity_search", entity_search_tool, ENTITY_SEARCH_SCHEMA)
tool_registry.register_builtin_tool("learning_path", learning_path_tool, LEARNING_PATH_SCHEMA)
tool_registry.register_builtin_tool("image_ocr", image_ocr_tool, IMAGE_OCR_SCHEMA)
tool_registry.register_builtin_tool("image_vision", image_vision_tool, IMAGE_VISION_SCHEMA)
tool_registry.register_builtin_tool("speech_to_text", speech_to_text_tool, SPEECH_TO_TEXT_SCHEMA)
tool_registry.register_builtin_tool("text_to_speech", text_to_speech_tool, TEXT_TO_SPEECH_SCHEMA)
_registered = True
n = tool_registry.builtin_tool_count()
names = tool_registry.builtin_tool_names()
pid = os.getpid()
if n < _EXPECTED_BUILTIN:
logger.warning(
"内置工具注册数量异常: pid=%s count=%s 期望>=%s names=%sLLM 工具调用可能失效)",
pid,
n,
_EXPECTED_BUILTIN,
names,
)
else:
logger.info(
"内置工具就绪 pid=%s count=%s names=%sCelery Worker 若缺此项日志,说明未加载 workflow_tasks / 未执行 bootstrap",
pid,
n,
names,
)