Create a chatbot that can ingest files

This tutorial shows you how to build a chatbot that can read and answer questions about files you upload, such as meeting notes or job applications.

For example, you could upload a contract and ask, “What are the termination clauses in this agreement?” Or upload a resume and ask, “Does this candidate have experience with marketing analytics?”

The main focus of this tutorial is to show you how to provide files as input to a Langflow flow, so your chatbot can use the content of those files in its responses.

Prerequisites

• Install and start Langflow
• Create a Langflow API key
• Create an OpenAI API key

This tutorial uses an OpenAI LLM. If you want to use a different provider, you need a valid credential for that provider.

Create a flow that accepts file input

To ingest files, your flow must have a File component attached to a component that receives input, such as a Prompt Template or Agent component.

The following steps modify the Basic Prompting template to accept file input:

1. In Langflow, click New Flow, and then select the Basic Prompting template.

2. In the Language Model component, enter your OpenAI API key.

   If you want to use a different provider or model, edit the Model Provider, Model Name, and API Key fields accordingly.

3. To verify that your API key is valid, click Playground, and then ask the LLM a question. The LLM should respond according to the specifications in the Prompt Template component's Template field.

4. Exit the Playground, and then modify the Prompt Template component to accept file input in addition to chat input. To do this, edit the Template field, and then replace the default prompt with the following text:

   
   

   _10

   ChatInput:

   _10

   {chat-input}

   _10

   File:

   _10

   {file}

   
   
   tip

   You can use any string to name your template variables. These strings become the names of the fields (input ports) on the Prompt Template component.

   For this tutorial, the variables are named after the components that connect to them: chat-input for the Chat Input component and file for the File component.

5. Add a File component to the flow, and then connect the Raw Content output port to the Prompt Template component's file input port. To connect ports, click and drag from one port to the other.

   You can add files directly to the File component to pre-load input before running the flow, or you can load files at runtime. The next section of this tutorial covers runtime file uploads.

   At this point your flow has five components. The Chat Input and File components are connected to the Prompt Template component's input ports. Then, the Prompt Template component's output port is connected to the Language Model component's input port. Finally, the Language Model component's output port is connected to the Chat Output component, which returns the final response to the user.

   

Send requests to your flow from a Python application

This section of the tutorial demonstrates how you can send file input to a flow from an application.

To do this, your application must send a POST /run request to your Langflow server with the file you want to upload and a text prompt. The result includes the outcome of the flow run and the LLM's response.

This example uses a local Langflow instance, and it asks the LLM to evaluate a sample resume. If you don't have a resume on hand, you can download fake-resume.txt.

tip

For help with constructing file upload requests in Python, JavaScript, and curl, see the Langflow File Upload Utility.

1. To construct the request, gather the following information:

   • LANGFLOW_SERVER_ADDRESS: Your Langflow server's domain. The default value is 127.0.0.1:7860. You can get this value from the code snippets on your flow's API access pane.
   • FLOW_ID: Your flow's UUID or custom endpoint name. You can get this value from the code snippets on your flow's API access pane.
   • FILE_COMPONENT_ID: The UUID of the File component in your flow, such as File-KZP68. To find the component ID, open your flow in Langflow, click the File component, and then click Controls. The component ID is at the top of the Controls pane.
   • CHAT_INPUT: The message you want to send to the Chat Input of your flow, such as Evaluate this resume for a job opening in my Marketing department.
   • FILE_NAME and FILE_PATH: The name and path to the local file that you want to send to your flow.
   • LANGFLOW_API_KEY: A valid Langflow API key.

2. Copy the following script into a Python file, and then replace the placeholders with the information you gathered in the previous step:

   
   

   _51

   # Python example using requests

   _51

   import requests

   _51

   import json

   _51

   _51

   # 1. Set the upload URL

   _51

   url = "http://LANGFLOW_SERVER_ADDRESS/api/v2/files/"

   _51

   _51

   # 2. Prepare the file and payload

   _51

   payload = {}

   _51

   files = [

   _51

   ('file', ('FILE_PATH', open('FILE_NAME', 'rb'), 'application/octet-stream'))

   _51

   ]

   _51

   headers = {

   _51

   'Accept': 'application/json',

   _51

   'x-api-key': 'LANGFLOW_API_KEY'

   _51

   }

   _51

   _51

   # 3. Upload the file to Langflow

   _51

   response = requests.request("POST", url, headers=headers, data=payload, files=files)

   _51

   print(response.text)

   _51

   _51

   # 4. Get the uploaded file path from the response

   _51

   uploaded_data = response.json()

   _51

   uploaded_path = uploaded_data.get('path')

   _51

   _51

   # 5. Call the Langflow run endpoint with the uploaded file path

   _51

   run_url = "http://LANGFLOW_SERVER_ADDRESS/api/v1/run/FLOW_ID"

   _51

   run_payload = {

   _51

   "input_value": "CHAT_INPUT",

   _51

   "output_type": "chat",

   _51

   "input_type": "chat",

   _51

   "tweaks": {

   _51

   "FILE_COMPONENT_ID": {

   _51

   "path": uploaded_path

   _51

   }

   _51

   }

   _51

   }

   _51

   run_headers = {

   _51

   'Content-Type': 'application/json',

   _51

   'Accept': 'application/json',

   _51

   'x-api-key': 'LANGFLOW_API_KEY'

   _51

   }

   _51

   run_response = requests.post(run_url, headers=run_headers, data=json.dumps(run_payload))

   _51

   langflow_data = run_response.json()

   _51

   # Output only the message

   _51

   message = None

   _51

   try:

   _51

   message = langflow_data['outputs'][0]['outputs'][0]['results']['message']['data']['text']

   _51

   except (KeyError, IndexError, TypeError):

   _51

   pass

   _51

   print(message)

   
   

   This script contains two requests.

   The first request uploads a file, such as fake-resume.txt, to your Langflow server at the /v2/files endpoint. This request returns a file path that can be referenced in subsequent Langflow requests, such as 02791d46-812f-4988-ab1c-7c430214f8d5/fake-resume.txt

   The second request sends a chat message to the Langflow flow at the /v1/run/ endpoint. The tweaks parameter includes the path to the uploaded file as the variable uploaded_path, and sends this file directly to the File component.

3. Save and run the script to send the requests and test the flow.

   The initial output contains the JSON response object from the file upload endpoint, including the internal path where Langflow stores the file. Then, the LLM retrieves the file and evaluates its content, in this case the suitability of the resume for a job position.

   Result

   The following is an example response from this tutorial's flow. Due to the nature of LLMs and variations in your inputs, your response might be different.

   
   

   _23

   {"id":"793ba3d8-5e7a-4499-8b89-d9a7b6325fee","name":"fake-resume (1)","path":"02791d46-812f-4988-ab1c-7c430214f8d5/fake-resume.txt","size":1779,"provider":null}

   _23

   The resume for Emily J. Wilson presents a strong candidate for a position in your Marketing department. Here are some key points to consider:

   _23

   _23

   ### Strengths:

   _23

   1. **Experience**: With over 8 years in marketing, Emily has held progressively responsible positions, culminating in her current role as Marketing Director. This indicates a solid foundation in the field.

   _23

   _23

   2. **Quantifiable Achievements**: The resume highlights specific accomplishments, such as a 25% increase in brand recognition and a 30% sales increase after launching new product lines. These metrics demonstrate her ability to drive results.

   _23

   _23

   3. **Diverse Skill Set**: Emily's skills encompass various aspects of marketing, including strategy development, team management, social media marketing, event planning, and data analysis. This versatility can be beneficial in a dynamic marketing environment.

   _23

   _23

   4. **Educational Background**: Her MBA and a Bachelor's degree in Marketing provide a strong academic foundation, which is often valued in marketing roles.

   _23

   _23

   5. **Certifications**: The Certified Marketing Professional (CMP) and Google Analytics Certification indicate a commitment to professional development and staying current with industry standards.

   _23

   _23

   ### Areas for Improvement:

   _23

   1. **Specificity in Skills**: While the skills listed are relevant, providing examples of how she has applied these skills in her previous roles could strengthen her resume further.

   _23

   _23

   2. **References**: While stating that references are available upon request is standard, including a couple of testimonials or notable endorsements could enhance credibility.

   _23

   _23

   3. **Formatting**: Ensure that the resume is visually appealing and easy to read. Clear headings and bullet points help in quickly identifying key information.

   _23

   _23

   ### Conclusion:

   _23

   Overall, Emily J. Wilson's resume reflects a well-rounded marketing professional with a proven track record of success. If her experience aligns with the specific needs of your Marketing department, she could be a valuable addition to your team. Consider inviting her for an interview to further assess her fit for the role.

   
   

Next steps

To continue building on this tutorial, try these next steps.

Process multiple files loaded at runtime

To process multiple files in a single flow run, add a separate File component for each file you want to ingest. Then, modify your script to upload each file, retrieve each returned file path, and then pass a unique file path to each File component ID.

For example, you can modify tweaks to accept multiple File components. The following code is just an example; it isn't working code:

_13

## set multiple file paths

_13

file_paths = {

_13

FILE_COMPONENT_1: uploaded_path_1,

_13

FILE_COMPONENT_2: uploaded_path_2

_13

}

_13

_13

def chat_with_flow(input_message, file_paths):

_13

"""Compare the contents of these two files."""

_13

run_url = f"{LANGFLOW_SERVER_ADDRESS}/api/v1/run/{FLOW_ID}"

_13

# Prepare tweaks with both file paths

_13

tweaks = {}

_13

for component_id, file_path in file_paths.items():

_13

tweaks[component_id] = {"path": file_path}

You can also use a Directory component to load all files in a directory or pass an archive file to the File component.

Upload external files at runtime

To upload files from another machine that isn't your local environment, your Langflow server must first be accessible over the internet. Then, authenticated users can upload files to your public Langflow server's /v2/files/ endpoint, as shown in the tutorial. For more information, see Langflow deployment overview.

Preload files outside the chat session

You can use the File component to load files anywhere in a flow, not just in a chat session.

In the visual editor, you can preload files to the File component by selecting them from your local machine or Langflow file management.

For example, you can preload an instructions file for a prompt template, or you can preload a vector store with documents that you want to query in a Retrieval Augmented Generation (RAG) flow.

For more information about the File component and other data loading components, see Data components.