Background

A simple tool that automates the qualification of a partner(reseller/integrator/editor) by finding its website, industries, business functions and services in a structured way.

• It provides browsing ability to reduce hallucination problem and improve the information precision.

• It provides single mode and batch mode so that you can integrate as an API or in a scheduler easily.

An example of qualification result:

• (industry column) Target your industry
• (job column) The jobs that can be targeted through their industry
• (offers column) The services they offer.

Install

pip install partenaire-qualif

Usage

The tool provides browsing ability which can be activated by setting the parameter browsing to True. Sample code below.

from qualif import CompanyQualificationTool
from openai import AzureOpenAI, OpenAI
import pandas as pd


# Necessary Variables
azure_endpoint = "https://..."  # azure openai endpoint
apikey = "..." # azure openai apikey
apiversion = "..." # azure openai api version
model = "..." # azure openai model name

# Setup OpenAI CLI
client = AzureOpenAI(
    azure_endpoint=azure_endpoint,
    api_key=apikey,
    api_version=apiversion,
)

# Use QualificationTool without browsing
browsing = False
tool = CompanyQualificationTool(client=client, model=model, temperature=0, browsing=browsing)
result = tool.qualify(partner_name="Your_Partner_Name")


# Use QualificationTool with browsing
browsing = True
browsing_key = '...' # azure bing search apikey
tool = CompanyQualificationTool(client=client, model=model, temperature=0, browsing=browsing, browsing_key=browsing_key)
result = tool.qualify(partner_name="Your_Partner_Name")
print(result)

The tool also allow to execute in batch mode by providing a panda dataframe. Sample code as below.

# Construct panda dataframe from a excel file that has a column containing the name of partners
df_partenaires = pd.read_excel("Partenaires.xlsx",sheet_name="Partenaires")

# Use the batch method to execute the qualification in batch
final_df = tool.batch(df_partners=df_partenaires, col_name_partner="Nom_Partenaire")
final_df.to_excel("Partenaires_qualified.xlsx")

Release

source release.sh {NEW_VERSION}  # example 1.1.1

Maintainers

@YMURONG

License