Skip to main content

FactSet

Conversational client library for Python

API Version PyPi Apache-2 license

This API lets you send a question to FactSet's conversational agent and get an answer back, formatted the same way OpenAI's Chat Completions API formats its responses. That means tools already built to talk to OpenAI-style chat APIs (chat UIs, IDE assistants, evaluation harnesses, and so on) can talk to conversational agent with little or no changes.

A few things work a bit differently from a typical OpenAI integration:

  • Even though you can send a full list of messages (a "chat history"), only the most recent message from the user is actually used as the question being asked. Any earlier messages, or a system prompt, are ignored.
  • Instead, the API keeps track of the conversation for you. Each conversation has an id (conversation_id). Send that id back on your next request and the assistant will remember what was said before and continue the same conversation. Leave it out and a new conversation is started, with a fresh id handed back to you to reuse next time.
  • The answer isn't returned all at once — it's streamed back a little at a time as it's generated, using a format called Server-Sent Events (text/event-stream). This is the same streaming format OpenAI's API uses, so most existing client libraries already know how to read it. The stream always ends with a data: [DONE] line so the client knows there's nothing more coming.

A note on FactSet's internal API standards: this endpoint's contract is intentionally OpenAI's Chat Completions API contract, not FactSet's own API design standards — that's the entire point of the endpoint, since it lets existing OpenAI-compatible clients work against it unmodified. Where the two disagree, OpenAI's contract wins by design, and the FactSet standard does not apply. In particular:

  • Member names (finish_reason, reasoning_content, factset_configuration, conversation_id, return_raw_link_context, factset_context) are snake_case, matching OpenAI's own field naming (and, for reasoning_content, DeepSeek's), rather than FactSet's camelCase convention.
  • Error responses are plain text, matching this service's shared error-handling framework, rather than FactSet's structured {\"errors\": [...]} JSON envelope.

This Python package is automatically generated by the OpenAPI Generator project:

  • API version: 0.2.0
  • SDK version: 0.0.1
  • Build package: org.openapitools.codegen.languages.PythonClientCodegen

For more information, please visit https://developer.factset.com/contact

Requirements

  • Python >= 3.7

Installation

Poetry

poetry add fds.sdk.utils fds.sdk.Conversational==0.0.1

pip

pip install fds.sdk.utils fds.sdk.Conversational==0.0.1

Usage

  1. Generate authentication credentials.
  2. Setup Python environment.
    1. Install and activate python 3.10+. If you're using pyenv:

      pyenv install 3.10.0
      pyenv shell 3.10.0
      
    2. (optional) Install poetry.

  3. Install dependencies.
  4. Run the following:

Example Code

from fds.sdk.utils.authentication import ConfidentialClient

import fds.sdk.Conversational
from fds.sdk.Conversational.api import chat_api
from fds.sdk.Conversational.models import *
from dateutil.parser import parse as dateutil_parser
from pprint import pprint

# See configuration.py for a list of all supported configuration parameters.

# Examples for each supported authentication method are below,
# choose one that satisfies your use case.

# (Preferred) OAuth 2.0: FactSetOAuth2
# See https://github.com/FactSet/enterprise-sdk#oauth-20
# for information on how to create the app-config.json file
#
# The confidential client instance should be reused in production environments.
# See https://github.com/FactSet/enterprise-sdk-utils-python#authentication
# for more information on using the ConfidentialClient class
configuration = fds.sdk.Conversational.Configuration(
    fds_oauth_client=ConfidentialClient('/path/to/app-config.json')
)

# Basic authentication: FactSetApiKey
# See https://github.com/FactSet/enterprise-sdk#api-key
# for information how to create an API key
# configuration = fds.sdk.Conversational.Configuration(
#     username='USERNAME-SERIAL',
#     password='API-KEY'
# )

# Enter a context with an instance of the API client
with fds.sdk.Conversational.ApiClient(configuration) as api_client:
    # Create an instance of the API class
    api_instance = chat_api.ChatApi(api_client)
    chat_completion_request = ChatCompletionRequest(
        model="fds-answer-2.0",
        stream=True,
        messages=[
            ChatMessage(
                role="system",
                content="content_example",
            ),
        ],
        factset_configuration=FactSetConfiguration(
            conversation_id="conversation_id_example",
            return_raw_link_context=False,
        ),
    ) # ChatCompletionRequest | 
    consumer = "developer-portal" # str | A short name identifying which application or integration is making this call, used for FactSet's own usage tracking. If omitted, usage is tracked under a generic default value rather than the request being rejected. (optional)

    try:
        # Ask a question and stream back the answer (OpenAI-compatible)
        # example passing only required values which don't have defaults set
        # and optional values
        api_response = api_instance.chat_completion(chat_completion_request, consumer=consumer)

        pprint(api_response)
    except fds.sdk.Conversational.ApiException as e:
        print("Exception when calling ChatApi->chat_completion: %s\n" % e)

    # # Get response, http status code and response headers
    # try:
    #     # Ask a question and stream back the answer (OpenAI-compatible)
    #     api_response, http_status_code, response_headers = api_instance.chat_completion_with_http_info(chat_completion_request, consumer=consumer)


    #     pprint(api_response)
    #     pprint(http_status_code)
    #     pprint(response_headers)
    # except fds.sdk.Conversational.ApiException as e:
    #     print("Exception when calling ChatApi->chat_completion: %s\n" % e)

    # # Get response asynchronous
    # try:
    #     # Ask a question and stream back the answer (OpenAI-compatible)
    #     async_result = api_instance.chat_completion_async(chat_completion_request, consumer=consumer)
    #     api_response = async_result.get()


    #     pprint(api_response)
    # except fds.sdk.Conversational.ApiException as e:
    #     print("Exception when calling ChatApi->chat_completion: %s\n" % e)

    # # Get response, http status code and response headers asynchronous
    # try:
    #     # Ask a question and stream back the answer (OpenAI-compatible)
    #     async_result = api_instance.chat_completion_with_http_info_async(chat_completion_request, consumer=consumer)
    #     api_response, http_status_code, response_headers = async_result.get()


    #     pprint(api_response)
    #     pprint(http_status_code)
    #     pprint(response_headers)
    # except fds.sdk.Conversational.ApiException as e:
    #     print("Exception when calling ChatApi->chat_completion: %s\n" % e)

Using Pandas

To convert an API response to a Pandas DataFrame, it is necessary to transform it first to a dictionary.

import pandas as pd

response_dict = api_response.to_dict()['data']

simple_json_response = pd.DataFrame(response_dict)
nested_json_response = pd.json_normalize(response_dict)

Debugging

The SDK uses the standard library logging module.

Setting debug to True on an instance of the Configuration class sets the log-level of related packages to DEBUG and enables additional logging in Pythons HTTP Client.

Note: This prints out sensitive information (e.g. the full request and response). Use with care.

import logging
import fds.sdk.Conversational

logging.basicConfig(level=logging.DEBUG)

configuration = fds.sdk.Conversational.Configuration(...)
configuration.debug = True

Configure a Proxy

You can pass proxy settings to the Configuration class:

  • proxy: The URL of the proxy to use.
  • proxy_headers: a dictionary to pass additional headers to the proxy (e.g. Proxy-Authorization).
import fds.sdk.Conversational

configuration = fds.sdk.Conversational.Configuration(
    # ...
    proxy="http://secret:password@localhost:5050",
    proxy_headers={
        "Custom-Proxy-Header": "Custom-Proxy-Header-Value"
    }
)

Custom SSL Certificate

TLS/SSL certificate verification can be configured with the following Configuration parameters:

  • ssl_ca_cert: a path to the certificate to use for verification in PEM format.
  • verify_ssl: setting this to False disables the verification of certificates. Disabling the verification is not recommended, but it might be useful during local development or testing.
import fds.sdk.Conversational

configuration = fds.sdk.Conversational.Configuration(
    # ...
    ssl_ca_cert='/path/to/ca.pem'
)

Request Retries

In case the request retry behaviour should be customized, it is possible to pass a urllib3.Retry object to the retry property of the Configuration.

from urllib3 import Retry
import fds.sdk.Conversational

configuration = fds.sdk.Conversational.Configuration(
    # ...
)

configuration.retries = Retry(total=3, status_forcelist=[500, 502, 503, 504])

Documentation for API Endpoints

All URIs are relative to https://api.factset.com/conversational/v0

Class Method HTTP request Description
ChatApi chat_completion POST /chat/completions Ask a question and stream back the answer (OpenAI-compatible)
ChatApi list_models GET /models List the models you can use (OpenAI-compatible)

Documentation For Models

Documentation For Authorization

FactSetApiKey

  • Type: HTTP basic authentication

FactSetOAuth2

  • Type: OAuth
  • Flow: application
  • Authorization URL:
  • Scopes: N/A

Notes for Large OpenAPI documents

If the OpenAPI document is large, imports in fds.sdk.Conversational.apis and fds.sdk.Conversational.models may fail with a RecursionError indicating the maximum recursion limit has been exceeded. In that case, there are a couple of solutions:

Solution 1: Use specific imports for apis and models like:

  • from fds.sdk.Conversational.api.default_api import DefaultApi
  • from fds.sdk.Conversational.model.pet import Pet

Solution 2: Before importing the package, adjust the maximum recursion limit as shown below:

import sys
sys.setrecursionlimit(1500)
import fds.sdk.Conversational
from fds.sdk.Conversational.apis import *
from fds.sdk.Conversational.models import *

Contributing

Please refer to the contributing guide.

Copyright 2026 FactSet Research Systems Inc

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Metadata

Release files for fds.sdk.Conversational 0.0.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for fds.sdk.Conversational 0.0.1
File Size Uploaded Provenance Actions
fds_sdk_conversational-0.0.1.tar.gz 60.3 kB Without provenance Details for fds_sdk_conversational-0.0.1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for fds.sdk.Conversational 0.0.1
File Interpreter ABI Platform Actions
fds_sdk_conversational-0.0.1-py3-none-any.whl Python 3 none any Details for fds_sdk_conversational-0.0.1-py3-none-any.whl

Total release size: 149.8 kB

Release files / fds_sdk_conversational-0.0.1.tar.gz

Download URL fds_sdk_conversational-0.0.1.tar.gz
Size 60.3 kB
Tags Source
SHA-256 checksum
How to use checksums
f6fe182c0f59deacbb1004fa0bf83dfab7184a74363f6ab4e8deb74eb0ecb3db
BLAKE2b-256 checksum
How to use checksums
29cb46199d9f14262ee26ed4f2adcd0f65161c3e7229a251605bb5d3f4cdc768
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.14.9

Release files / fds_sdk_conversational-0.0.1-py3-none-any.whl

Download URL fds_sdk_conversational-0.0.1-py3-none-any.whl
Size 89.4 kB
Tags Python 3
SHA-256 checksum
How to use checksums
6964e4592f61ba448c3250009033d95d828c0a57109b4dee6c731d6739668633
BLAKE2b-256 checksum
How to use checksums
dcfea712fdfc605d9780ae2c1091f01c7072046348b38c0665b9c8cfd6840dad
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.14.9

Security for fds.sdk.Conversational 0.0.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.0.1 This release

2 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page