my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

#1	---
#2	title: '💬 chat'
#3	---
#4
#5	`chat()` method allows you to chat over your data sources using a user-friendly chat API. You can find the signature below:
#6
#7	### Parameters
#8
#9	<ParamField path="input_query" type="str">
#10	Question to ask
#11	</ParamField>
#12	<ParamField path="config" type="BaseLlmConfig" optional>
#13	Configure different llm settings such as prompt, temprature, number_documents etc.
#14	</ParamField>
#15	<ParamField path="dry_run" type="bool" optional>
#16	The purpose is to test the prompt structure without actually running LLM inference. Defaults to `False`
#17	</ParamField>
#18	<ParamField path="where" type="dict" optional>
#19	A dictionary of key-value pairs to filter the chunks from the vector database. Defaults to `None`
#20	</ParamField>
#21	<ParamField path="session_id" type="str" optional>
#22	Session ID of the chat. This can be used to maintain chat history of different user sessions. Default value: `default`
#23	</ParamField>
#24	<ParamField path="citations" type="bool" optional>
#25	Return citations along with the LLM answer. Defaults to `False`
#26	</ParamField>
#27
#28	### Returns
#29
#30	<ResponseField name="answer" type="str \| tuple">
#31	If `citations=False`, return a stringified answer to the question asked. <br />
#32	If `citations=True`, returns a tuple with answer and citations respectively.
#33	</ResponseField>
#34
#35	## Usage
#36
#37	### With citations
#38
#39	If you want to get the answer to question and return both answer and citations, use the following code snippet:
#40
#41	```python With Citations
#42	from embedchain import App
#43
#44	# Initialize app
#45	app = App()
#46
#47	# Add data source
#48	app.add("https://www.forbes.com/profile/elon-musk")
#49
#50	# Get relevant answer for your query
#51	answer, sources = app.chat("What is the net worth of Elon?", citations=True)
#52	print(answer)
#53	# Answer: The net worth of Elon Musk is $221.9 billion.
#54
#55	print(sources)
#56	# [
#57	# (
#58	# 'Elon Musk PROFILEElon MuskCEO, Tesla$247.1B$2.3B (0.96%)Real Time Net Worthas of 12/7/23 ...',
#59	# {
#60	# 'url': 'https://www.forbes.com/profile/elon-musk',
#61	# 'score': 0.89,
#62	# ...
#63	# }
#64	# ),
#65	# (
#66	# '74% of the company, which is now called X.Wealth HistoryHOVER TO REVEAL NET WORTH BY YEARForbes ...',
#67	# {
#68	# 'url': 'https://www.forbes.com/profile/elon-musk',
#69	# 'score': 0.81,
#70	# ...
#71	# }
#72	# ),
#73	# (
#74	# 'founded in 2002, is worth nearly $150 billion after a $750 million tender offer in June 2023 ...',
#75	# {
#76	# 'url': 'https://www.forbes.com/profile/elon-musk',
#77	# 'score': 0.73,
#78	# ...
#79	# }
#80	# )
#81	# ]
#82	```
#83
#84	<Note>
#85	When `citations=True`, note that the returned `sources` are a list of tuples where each tuple has two elements (in the following order):
#86	1. source chunk
#87	2. dictionary with metadata about the source chunk
#88	- `url`: url of the source
#89	- `doc_id`: document id (used for book keeping purposes)
#90	- `score`: score of the source chunk with respect to the question
#91	- other metadata you might have added at the time of adding the source
#92	</Note>
#93
#94
#95	### Without citations
#96
#97	If you just want to return answers and don't want to return citations, you can use the following example:
#98
#99	```python Without Citations
#100	from embedchain import App
#101
#102	# Initialize app
#103	app = App()
#104
#105	# Add data source
#106	app.add("https://www.forbes.com/profile/elon-musk")
#107
#108	# Chat on your data using `.chat()`
#109	answer = app.chat("What is the net worth of Elon?")
#110	print(answer)
#111	# Answer: The net worth of Elon Musk is $221.9 billion.
#112	```
#113
#114	### With session id
#115
#116	If you want to maintain chat sessions for different users, you can simply pass the `session_id` keyword argument. See the example below:
#117
#118	```python With session id
#119	from embedchain import App
#120
#121	app = App()
#122	app.add("https://www.forbes.com/profile/elon-musk")
#123
#124	# Chat on your data using `.chat()`
#125	app.chat("What is the net worth of Elon Musk?", session_id="user1")
#126	# 'The net worth of Elon Musk is $250.8 billion.'
#127	app.chat("What is the net worth of Bill Gates?", session_id="user2")
#128	# "I don't know the current net worth of Bill Gates."
#129	app.chat("What was my last question", session_id="user1")
#130	# 'Your last question was "What is the net worth of Elon Musk?"'
#131	```
#132
#133	### With custom context window
#134
#135	If you want to customize the context window that you want to use during chat (default context window is 3 document chunks), you can do using the following code snippet:
#136
#137	```python with custom chunks size
#138	from embedchain import App
#139	from embedchain.config import BaseLlmConfig
#140
#141	app = App()
#142	app.add("https://www.forbes.com/profile/elon-musk")
#143
#144	query_config = BaseLlmConfig(number_documents=5)
#145	app.chat("What is the net worth of Elon Musk?", config=query_config)
#146	```
#147
#148	### With Mem0 to store chat history
#149
#150	Mem0 is a cutting-edge long-term memory for LLMs to enable personalization for the GenAI stack. It enables LLMs to remember past interactions and provide more personalized responses.
#151
#152	In order to use Mem0 to enable memory for personalization in your apps:
#153	- Install the [`mem0`](https://docs.mem0.ai/) package using `pip install mem0ai`.
#154	- Prepare config for `memory`, refer [Configurations](docs/api-reference/advanced/configuration.mdx).
#155
#156	```python with mem0
#157	from embedchain import App
#158
#159	config = {
#160	"memory": {
#161	"top_k": 5
#162	}
#163	}
#164
#165	app = App.from_config(config=config)
#166	app.add("https://www.forbes.com/profile/elon-musk")
#167
#168	app.chat("What is the net worth of Elon Musk?")
#169	```
#170
#171	## How Mem0 works:
#172	- Mem0 saves context derived from each user question into its memory.
#173	- When a user poses a new question, Mem0 retrieves relevant previous memories.
#174	- The `top_k` parameter in the memory configuration specifies the number of top memories to consider during retrieval.
#175	- Mem0 generates the final response by integrating the user's question, context from the data source, and the relevant memories.
#176

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public