CoolFace
Apppublic

Raptive-NC/Creator_Catalog

sourceHugging Faceupdated 4mo agoView on Hugging Face
0likes
App README

Creator Catalog Dashboard

This Streamlit dashboard supports a campaign-first workflow:

  1. 1.Filter selection
  2. 2.Interactive chat to refine creator suggestions
  3. 3.Finalize creators
  4. 4.Generate a slide deck (.pptx)

Key behavior

  • —Uses Snowflake as source of truth:
  • —ANALYTICS.SIGMA_SCRATCH.CREATOR_CATALOG_FULL_DATA_SET
  • —ANALYTICS.ADTHRIVE.SITE_EXTENDED
  • —Joins CREATOR_CATALOG_FULL_DATA_SET.SITE_NAME = SITE_EXTENDED.SITE_NAME.
  • —Applies hard filters at query time:
  • —STATUS = Active
  • —IG_AUDIENCE_TIER NOT IN ('Micro', 'Nano')
  • —Uses an advertiser input to:
  • —exclude creators with matching Brand Avoidance, and
  • —help prioritize relevant creators.
  • —Treats blank or N/A advertiser input as no avoidance constraint.
  • —Prioritizes Raptive Selects when enabled.
  • —Requires Instagram account matching when Instagram is selected.
  • —Produces 10 creator candidates and allows conversational refinement.

Dashboard workflow instructions

  1. 1.Select campaign filters (Advertiser, Content, Platform, Tier) and write a campaign Description.
  2. 2.Click Start Creator Chat to generate the first set of creator recommendations.
  3. 3.Use chat follow-ups to refine the list (example: keep creators 1-4, replace creators 5-6).
  4. 4.Click Finalize Creators when recommendations are approved.
  5. 5.Click Generate Slide to export a .pptx.
  6. 6.Optional: place creator headshots in src/headshots using a site-name or creator-name slug as filename.

Required environment variables

  • —OPENAI_API_KEY (or REPO_SECRET_OPENAI_API_KEY)
  • —Snowflake key-pair auth:
  • —SNOWFLAKE_ACCOUNT
  • —SNOWFLAKE_USER
  • —SNOWFLAKE_ROLE
  • —SNOWFLAKE_WAREHOUSE
  • —SNOWFLAKE_DATABASE
  • —SNOWFLAKE_SCHEMA
  • —SNOWFLAKE_PRIVATE_KEY
  • —optional SNOWFLAKE_PRIVATE_KEY_PASSPHRASE
  • —Optional:
  • —OPENAI_BASE_URL
  • —CUSTOM_GPT_MODEL
  • —CUSTOM_GPT_INSTRUCTIONS

Local development

  1. 1.Install dependencies from requirements.txt.
  2. 2.Update .env placeholders in the project root.
  3. 3.Load env vars in your shell:
bash
   set -a
   source .env
   set +a
  1. 1.Run:
bash
   streamlit run src/streamlit_app.py