Robot Vacuum Agent

https://github.com/zhdhnsfdx/rag-agent
Admin
This project is a Python-based RAG (Retrieval Augmented Generation) system designed to answer...
Tokens:51,514
Snippets:793
Trust Score:2.9
Update:4 months ago
Show doc for...
Tokens:
Raw
### Runtime Context Example

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/ARCHITECTURE.md

Demonstrates the structure of the runtime context dictionary used for state-based behavior.

```python
context = {
    "report": False  # Set to True by monitor_tool when fill_context_for_report called
}
```

--------------------------------

### Agent Tool Integration Example

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/agent-tools.md

Illustrates how various tools, including fill_context_for_report, are registered and passed to the LangChain agent within the ReactAgent initialization.

```python
from agent.tools.agent_tools import (
    rag_summarize,
    get_weather,
    get_user_location,
    get_user_id,
    get_current_month,
    fetch_external_data,
    fill_context_for_report
)

tools=[
    rag_summarize,
    get_weather,
    get_user_location,
    get_user_id,
    get_current_month,
    fetch_external_data,
    fill_context_for_report
]
```

--------------------------------

### Initialize Default and Custom Loggers

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/logger-handler.md

Demonstrates how to initialize a default logger and a custom logger with specific configurations for console and file logging levels and a custom log file path. Includes example log messages.

```python
from utils.logger_handler import get_logger
import logging

# Default logger with INFO console and DEBUG file
logger = get_logger()

# Custom logger with DEBUG on both
custom_logger = get_logger(
    name="rag_service",
    console_level=logging.DEBUG,
    file_level=logging.DEBUG,
    log_file="logs/custom.log"
)

logger.info("Informational message")
logger.debug("Detailed debug message")
logger.error("Error message")
```

--------------------------------

### Load System Prompts

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Loads system prompts from the main prompt path. This is part of the initial setup for agent prompts.

```python
system_prompt = load_system_prompts()      # From main_prompt_path
rag_prompt = load_rag_prompts()            # From rag_summarize_prompt_path
report_prompt = load_report_prompts()      # From report_prompt_path
```

--------------------------------

### Runtime Configuration Modification Example

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Demonstrates modifying a configuration value at runtime. Note that this change does not affect services already instantiated with the old configuration.

```python
from utils.config_handler import chroma_conf

# Not recommended but possible
chroma_conf["k"] = 5  # Change retrieval count

# However, this doesn't affect already-instantiated services
# VectorStoreService already created with k=3
```

--------------------------------

### Get Weather Tool

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/agent-tools.md

Returns weather information for a specified city. Use this tool to get a mock weather forecast for a given city.

```python
@tool(description="获取指定城市的天气，以消息字符串的形式返回")
def get_weather(city: str) -> str
```

```python
weather = get_weather("深圳")
# Returns: "城市深圳天气为晴天，气温26摄氏度，空气湿度50%，南风1级，AQI21，最近6小时降雨概率极低"
```

--------------------------------

### Agent Configuration YAML Example

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Defines the path to external data files used by the agent. This setting is crucial for data loading operations.

```yaml
external_data_path: data/external/records.csv
```

--------------------------------

### Example Logger Output

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/logger-handler.md

Illustrates the typical output format for log messages, showing timestamp, logger name, level, source file, line number, and the message content.

```text
2025-06-16 14:23:45,123 - agent - INFO - react_agent.py:28 - Agent initialized
2025-06-16 14:23:46,456 - agent - ERROR - rag_service.py:87 - Failed to load documents
```

--------------------------------

### Full Document Indexing Workflow

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/file-handler.md

Illustrates the complete workflow for processing and indexing files, including listing, checking for duplicates, and loading documents. Ensure necessary setup like file listing and duplicate checking functions are available.

```python
# Full workflow:
allowed_files = listdir_with_allowed_type("data", (".txt", ".pdf"))
for path in allowed_files:
    md5 = get_file_md5_hex(path)
    if not already_indexed(md5):
        docs = get_file_documents(path)
        # ... chunk and index documents
```

--------------------------------

### Get and Use Logger

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/INDEX.md

Access the pre-configured logger instance for logging informational messages and errors.

```python
from utils.logger_handler import logger

logger.info("Message")
logger.error("Error occurred")
```

--------------------------------

### Example Error Log Messages

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Illustrates typical error log messages generated by the prompt loading functions, indicating configuration issues or file access problems.

```text
[load_system_prompts]在yaml配置项中没有main_prompt_path配置项
[load_system_prompts]解析系统提示词出错，[Errno 2] No such file or directory: '...'
```

--------------------------------

### Load Prompts Configuration

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/config-handler.md

Loads prompt file paths from a YAML configuration. Use this to get the paths for system, RAG, and report generation prompts.

```python
def load_prompts_config(
    config_path: str = get_abs_path("config/prompts.yml"),
    encoding: str = "utf-8"
) -> dict:
    # ... function body omitted for brevity
```

```python
prompts = load_prompts_config()
print(prompts["main_prompt_path"])  # "prompts/main_prompt.txt"
```

--------------------------------

### Get Documents from Files

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/file-handler.md

Determines the appropriate loader based on file extension and returns documents. Used within VectorStoreService for loading content.

```python
def get_file_documents(read_path: str):
    if read_path.endswith("txt"):
        return txt_loader(read_path)
    if read_path.endswith("pdf"):
        return pdf_loader(read_path)
    return []
```

--------------------------------

### CSV Data Format Example

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Illustrates the expected format for CSV files containing user usage data. Values are quoted and stripped during parsing.

```csv
user_id,feature,efficiency,consumables,comparison,month
"1001","feature_val","efficiency_val","consumables_val","comparison_val","2025-01"
```

--------------------------------

### Initializing Vector Store Service with Chroma Configuration

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Shows how to initialize the `VectorStoreService` which utilizes Chroma configuration for collection name, persistence, retrieval (k), and text splitting parameters.

```python
from utils.config_handler import chroma_conf
from rag.vector_store import VectorStoreService

vs = VectorStoreService()
# Uses chroma_conf for initialization:
# - Creates Chroma with collection_name and persist_directory
# - Configures retriever with k=3
# - Configures text splitter with chunk_size, chunk_overlap, separators

retriever = vs.get_retriever()
docs = retriever.invoke("query")  # Returns k=3 documents
```

--------------------------------

### Module-Specific Logger Configuration

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/logger-handler.md

Configure a logger for a specific module to manage its logging independently. This example shows how to get a module-specific logger and set both console and file logging levels to DEBUG. Import get_logger from utils.logger_handler.

```python
from utils.logger_handler import get_logger

logger = get_logger(
    name="my_module",
    console_level=logging.DEBUG,
    file_level=logging.DEBUG
)

logger.debug("Verbose tracing enabled for this module")
```

--------------------------------

### Get Current Month Tool

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/agent-tools.md

Returns a random month in YYYY-MM format. Use this tool to get a randomly selected month.

```python
@tool(description="获取当前月份，以纯字符串形式返回")
def get_current_month() -> str
```

```python
month = get_current_month()
print(month)  # Returns one of the 2025 months randomly
```

--------------------------------

### Get User ID Tool

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/agent-tools.md

Returns a random user ID. Use this tool to get a randomly selected user ID.

```python
@tool(description="获取用户的ID，以纯字符串形式返回")
def get_user_id() -> str
```

```python
user_id = get_user_id()
print(user_id)  # Returns one of the 10 user IDs randomly
```

--------------------------------

### Agent Initialization Usage

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Demonstrates how the prompt loading functions are used during agent initialization, where `load_system_prompts()` is called internally.

```python
from agent.react_agent import ReactAgent

agent = ReactAgent()
# Internally calls load_system_prompts() and sets it as system_prompt
```

--------------------------------

### Initialize VectorStoreService

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/VectorStoreService.md

Instantiate the VectorStoreService. This initializes the Chroma vector store and text splitter based on the configuration.

```python
vector_store = VectorStoreService()
```

--------------------------------

### Get User Location Tool

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/agent-tools.md

Returns a random user location city name. Use this tool to get a randomly selected city name for the user.

```python
@tool(description="获取用户所在城市的名称，以纯字符串形式返回")
def get_user_location() -> str
```

```python
location = get_user_location()
print(location)  # Randomly returns "深圳", "合肥", or "杭州"
```

--------------------------------

### Initialize RagSummarizeService

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/RagSummarizeService.md

Instantiate the RagSummarizeService. This initializes the vector store, retriever, prompt, model, and processing chain.

```python
rag = RagSummarizeService()
```

--------------------------------

### Configuration Hierarchy

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/ARCHITECTURE.md

Shows how configuration settings are loaded and applied throughout the project.

```text
YAML Files (config/*.yml)
    ↓
config_handler (load_*_config functions)
    ↓
Module-level instances (rag_conf, chroma_conf, etc.)
    ↓
Consumers (model.factory, rag.vector_store, etc.)
```

--------------------------------

### Instantiate ReactAgent

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/ReactAgent.md

Create an instance of the ReactAgent. This initializes the agent with pre-configured models, prompts, tools, and middleware.

```python
agent = ReactAgent()
```

--------------------------------

### Loading System Prompts

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Illustrates how to import and use functions for loading different types of system prompts, such as main, RAG summarize, and report prompts.

```python
from utils.prompt_loader import load_system_prompts, load_rag_prompts, load_report_prompts
```

--------------------------------

### Reading Multiple Configuration Files

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/configuration.md

Shows how to import and access various configuration objects like rag_conf, chroma_conf, prompts_conf, and agent_conf. Configuration values are accessed like dictionary keys.

```python
from utils.config_handler import rag_conf, chroma_conf, prompts_conf, agent_conf

# Access as dictionary
model_name = rag_conf["chat_model_name"]
k_value = chroma_conf["k"]
prompt_file = prompts_conf["main_prompt_path"]
data_file = agent_conf["external_data_path"]
```

--------------------------------

### Direct Model Access for Generation and Vectorization

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/model-factory.md

Provides examples of directly using the module-level chat_model for invoking chat and embed_model for text vectorization.

```python
from model.factory import chat_model, embed_model

# Chat model for generation
response = chat_model.invoke([
    {"role": "user", "content": "Explain RAG"}
])

# Embedding model for vectorization
query_embedding = embed_model.embed_query("robot vacuum reviews")
```

--------------------------------

### Initialize VectorStoreService

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/INDEX.md

Instantiate VectorStoreService for managing document loading and indexing. Documents are loaded from the 'data/' directory by default.

```python
from rag.vector_store import VectorStoreService

vs = VectorStoreService()
vs.load_document()  # Load documents from data/ directory
```

--------------------------------

### Safe Configuration Access with Error Handling

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/config-handler.md

Shows how to safely access configuration values using a try-except block to handle potential KeyErrors.

```python
from utils.config_handler import prompts_conf

try:
    path = prompts_conf["main_prompt_path"]
except KeyError:
    print("Configuration key not found")
```

--------------------------------

### Get Retriever Instance

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/VectorStoreService.md

Obtain a configured retriever for performing semantic searches. The retriever is set to fetch the top 3 most similar documents by default.

```python
vector_store = VectorStoreService()
retriever = vector_store.get_retriever()
```

--------------------------------

### Perform Semantic Search with Retriever

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/VectorStoreService.md

Use the retriever to find documents semantically related to a query. This example searches for 'robot vacuum small apartment'.

```python
vector_store = VectorStoreService()
retriever = vector_store.get_retriever()

# Use retriever to search
docs = retriever.invoke("robot vacuum small apartment")
for doc in docs:
    print(doc.page_content)
```

--------------------------------

### Access Pre-loaded Configurations

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/config-handler.md

Demonstrates how to access values from pre-loaded configuration dictionaries for chat models and chunk sizes.

```python
print(rag_conf["chat_model_name"])
print(chroma_conf["chunk_size"])
```

--------------------------------

### Get Logger Function Signature

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/logger-handler.md

The `get_logger` function creates and configures a logger with customizable console and file output levels, and an optional log file path.

```python
def get_logger(
    name: str = "agent",
    console_level: int = logging.INFO,
    file_level: int = logging.DEBUG,
    log_file: str | None = None
) -> logging.Logger
```

--------------------------------

### Initialize ReactAgent

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/INDEX.md

Instantiate the ReactAgent and stream its execution output. Useful for interactive agent tasks.

```python
from agent.react_agent import ReactAgent

agent = ReactAgent()
for chunk in agent.execute_stream("question"):
    print(chunk, end="", flush=True)
```

--------------------------------

### Get Project Root Directory

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/path-tool.md

Call get_project_root() to obtain the absolute path to the project's root directory. This function traverses up from the current module's location.

```python
from utils.path_tool import get_project_root

root = get_project_root()
print(root)
# Output: /workspace/home/rag-agent
```

--------------------------------

### Import Utility Functions

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/INDEX.md

Import functions for prompt loading, file handling, path resolution, and logging.

```python
from utils.prompt_loader import load_system_prompts, load_rag_prompts, load_report_prompts
from utils.file_handler import get_file_md5_hex, listdir_with_allowed_type, pdf_loader, txt_loader
from utils.path_tool import get_project_root, get_abs_path
from utils.logger_handler import get_logger, logger
```

--------------------------------

### Rag Summarize Tool

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/agent-tools.md

Retrieves and summarizes reference materials from the vector store using RAG. Use this tool to get a summarized answer with references for a given query.

```python
@tool(description="从向量存储中检索参考资料")
def rag_summarize(query: str) -> str
```

```python
result = rag_summarize("recommended equipment for facility maintenance")
print(result)  # Returns synthesized answer with references
```

--------------------------------

### load_system_prompts

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Loads the main system prompt for the agent from a configuration-defined file path.

```APIDOC
## Function: load_system_prompts

### Description
Loads the main system prompt for the agent.

### Signature
```python
def load_system_prompts() -> str
```

### Parameters
This function takes no parameters. The prompt file path is determined by `prompts_conf["main_prompt_path"]`.

### Returns
- `str`: A string containing the system prompt text.

### Throws
- `KeyError`: If `main_prompt_path` key is missing from `prompts_conf`.
- `FileNotFoundError` or `IOError`: If the prompt file cannot be opened.

### Example
```python
from utils.prompt_loader import load_system_prompts

prompt = load_system_prompts()
print(f"System prompt length: {len(prompt)} characters")
```
```

--------------------------------

### VectorStoreService Constructor

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/VectorStoreService.md

Initializes the VectorStoreService, setting up the Chroma vector store and text splitter based on configuration. It connects to a Chroma instance, sets an embedding model, and defines text splitting parameters.

```APIDOC
## VectorStoreService()

### Description
Initializes the VectorStoreService, setting up the Chroma vector store and text splitter based on configuration. It connects to a Chroma instance, sets an embedding model, and defines text splitting parameters.

### Parameters
No parameters.

### Returns
VectorStoreService instance

### Example
```python
vector_store = VectorStoreService()
```
```

--------------------------------

### Initialize React Agent

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/README.md

Initializes the ReactAgent and executes a streaming query, printing each chunk as it arrives. Ensure the agent is properly initialized before use.

```python
from agent.react_agent import ReactAgent

agent = ReactAgent()
for chunk in agent.execute_stream("What are your hours?"):
    print(chunk, end="", flush=True)
```

--------------------------------

### Summarize Query with RagSummarizeService

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/RagSummarizeService.md

Utilize the rag_summarize method to get a contextualized summary. This method retrieves relevant documents and uses a language model to generate a response based on the query and retrieved context.

```python
rag = RagSummarizeService()
result = rag.rag_summarize("Which robot vacuums work in small spaces?")
print(result)
# Output: Detailed response synthesizing query with retrieved knowledge
```

--------------------------------

### ReactAgent Constructor

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/ReactAgent.md

Initializes a ReactAgent instance. It is configured with a pre-defined chat model, system prompt, a set of RAG and utility tools, and middleware for monitoring and prompt switching. No parameters are required for initialization.

```APIDOC
## ReactAgent()

### Description
Initializes a ReactAgent instance with pre-configured model, system prompt, tools, and middleware.

### Method
__init__

### Parameters
No parameters.

### Returns
ReactAgent: An instance of the ReactAgent.

### Example
```python
agent = ReactAgent()
```
```

--------------------------------

### Import Configuration Utilities

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/config-handler.md

Imports all necessary configuration loading functions and pre-loaded configuration objects from the utils.config_handler module.

```python
from utils.config_handler import (
    load_rag_config,
    load_chroma_config,
    load_prompts_config,
    load_agent_config,
    rag_conf,
    chroma_conf,
    prompts_conf,
    agent_conf
)
```

--------------------------------

### Import Prompt Loader Utilities

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Import the necessary functions for loading different types of prompts.

```python
from utils.prompt_loader import (
    load_system_prompts,
    load_rag_prompts,
    load_report_prompts
)
```

--------------------------------

### Typical Python Import Pattern

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/module-exports.md

This snippet shows the most common import statements used when initializing new code, covering agent, configuration, logging, path resolution, RAG services, and models.

```python
from agent.react_agent import ReactAgent

from utils.config_handler import rag_conf, chroma_conf, prompts_conf, agent_conf

from utils.logger_handler import logger

from utils.path_tool import get_abs_path

from rag.rag_service import RagSummarizeService
from rag.vector_store import VectorStoreService

from model.factory import chat_model, embed_model
```

--------------------------------

### Import VectorStoreService

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/VectorStoreService.md

Import the VectorStoreService class from the rag.vector_store module.

```python
from rag.vector_store import VectorStoreService
```

--------------------------------

### Instantiate ChatModelFactory and Generate Model

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/model-factory.md

Demonstrates how to create an instance of ChatModelFactory and use its generator method to obtain a chat model. The model is configured via rag.yml.

```python
from model.factory import ChatModelFactory

factory = ChatModelFactory()
model = factory.generator()
# Returns ChatTongyi(model="qwen3-max")
```

--------------------------------

### Import Agent and Tools

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/INDEX.md

Import necessary classes and functions for agent orchestration and tool usage.

```python
from agent.react_agent import ReactAgent
from agent.tools.agent_tools import rag_summarize, get_weather, get_user_location, ...
```

```python
from agent.tools.middleware import monitor_tool, log_before_model, report_prompt_switch
```

--------------------------------

### Module-Level Configuration Objects

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/config-handler.md

Provides pre-loaded configuration dictionaries for RAG, Chroma, prompts, and agent settings. These objects are imported and used throughout the codebase for consistent configuration access.

```python
rag_conf = load_rag_config()
chroma_conf = load_chroma_config()
prompts_conf = load_prompts_config()
agent_conf = load_agent_config()
```

```python
from utils.config_handler import rag_conf, chroma_conf
```

--------------------------------

### Import Configuration Handlers

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/INDEX.md

Import configuration objects and loading functions for various modules.

```python
from utils.config_handler import rag_conf, chroma_conf, prompts_conf, agent_conf
from utils.config_handler import load_rag_config, load_chroma_config, load_prompts_config, load_agent_config
```

--------------------------------

### RagSummarizeService Initialization with chat_model and embed_model

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/model-factory.md

Shows how RagSummarizeService depends on both chat_model and embed_model. Ensure both are imported.

```python
from model.factory import chat_model, embed_model
from rag.rag_service import RagSummarizeService

rag = RagSummarizeService()  # Uses both chat_model and embed_model
```

--------------------------------

### ReactAgent Initialization with chat_model

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/model-factory.md

Illustrates how ReactAgent utilizes the module-level chat_model internally. Ensure chat_model is imported.

```python
from model.factory import chat_model
from agent.react_agent import ReactAgent

agent = ReactAgent()  # Uses chat_model internally
```

--------------------------------

### Dynamic Prompt Switching Usage

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Shows how prompt loading functions are used in dynamic prompt switching scenarios, such as with the `report_prompt_switch` middleware.

```python
from agent.tools.middleware import report_prompt_switch

# report_prompt_switch middleware uses these functions:
# - load_system_prompts() when context["report"] = False
# - load_report_prompts() when context["report"] = True
```

--------------------------------

### Load RAG Summarization Prompt

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Loads the RAG-specific summarization prompt template. This template includes placeholders for user input and retrieved context. It is configured via 'rag_summarize_prompt_path'.

```python
def load_rag_prompts() -> str:
    # ... implementation details ...
    pass
```

```python
from utils.prompt_loader import load_rag_prompts

rag_prompt = load_rag_prompts()
# Use with LangChain PromptTemplate
from langchain_core.prompts import PromptTemplate

template = PromptTemplate.from_template(rag_prompt)
formatted = template.format(
    input="robot vacuum recommendations",
    context="【参考资料1】: ..."
)
```

--------------------------------

### Import RagSummarizeService

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/RagSummarizeService.md

Import the RagSummarizeService class from the rag.rag_service module.

```python
from rag.rag_service import RagSummarizeService
```

--------------------------------

### load_rag_prompts

Source: https://github.com/zhdhnsfdx/rag-agent/blob/main/_autodocs/api-reference/prompt-loader.md

Loads the RAG-specific summarization prompt template, which includes placeholders for user input and retrieved context.

```APIDOC
## Function: load_rag_prompts

### Description
Loads the RAG-specific summarization prompt template.

### Signature
```python
def load_rag_prompts() -> str
```

### Parameters
This function takes no parameters. The prompt file path is determined by `prompts_conf["rag_summarize_prompt_path"]`.

### Returns
- `str`: A string containing the RAG prompt template with placeholders for `{input}` and `{context}`.

### Template Variables
- `{input}`: User query
- `{context}`: Formatted reference documents from vector retrieval

### Throws
- `KeyError`: If `rag_summarize_prompt_path` key is missing from `prompts_conf`.
- `FileNotFoundError` or `IOError`: If the prompt file cannot be opened.

### Example
```python
from utils.prompt_loader import load_rag_prompts

rag_prompt = load_rag_prompts()
# Use with LangChain PromptTemplate
from langchain_core.prompts import PromptTemplate

template = PromptTemplate.from_template(rag_prompt)
formatted = template.format(
    input="robot vacuum recommendations",
    context="【参考资料1】: ..."
)
```
```
Tokens:
Raw