Skip to content

README.md

Latest commit

 

History

History
89 lines (61 loc) · 4.54 KB

README.md

File metadata and controls

89 lines (61 loc) · 4.54 KB
Raw

Cohere Azure Workshops

Hands-on labs for using Cohere Embed-V5-Pro and Cohere rerank-v4.0 via Microsoft Foundry to build semantic search and reranking applications.

Labs

Lab Notebook Description
lab-1a-embed-getting-started lab-1a-embed-getting-started.ipynb Introduction to text embeddings with Cohere on Azure
lab-1b-embed-business-graphs lab-1b-embed-business-graphs.ipynb Multimodal semantic search — index and query business graph images
lab-2-rerank lab-2-rerank-getting-started.ipynb Semantic reranking — improve search relevance with Cohere Rerank
lab-2-rerank (optional) optional-lab-rerank_structured_data.ipynb Reranking structured data
lab-3-parse lab-3-parse-getting-started.ipynb Document parsing — extract structured blocks from documents

All notebooks use cohere.ClientV2 and connect to Microsoft Foundry via the /providers/cohere endpoint path.

Setup

Option A — Local (Windows — Command Prompt or PowerShell)

git clone https://github.com/cohere-ai/cohere-azure-workshops.git
cd cohere-azure-workshops
pip install -r requirements.txt
copy .env.example .env
# edit .env with your Azure credentials

Lab 1b note (Windows): Open lab-1b-embed-business-graphs.ipynb directly from the lab-1-embed\ folder in VS Code so the Jupyter kernel's working directory is set to that folder. This ensures relative paths like ./dataset/ and ./chroma_db resolve correctly.

Lab 3 note (Windows): Open lab-3-parse-getting-started.ipynb directly from the lab-3-parse\ folder in VS Code so the Jupyter kernel's working directory is set to that folder. This ensures the relative path ./sample_inputs/ resolves correctly.

Option B — Local (macOS / Linux)

git clone https://github.com/cohere-ai/cohere-azure-workshops.git
cd cohere-azure-workshops
pip install -r requirements.txt
cp .env.example .env
# edit .env with your Azure credentials

Option C — GitHub Codespaces

  1. Click Code → Codespaces → Create codespace on main
  2. Wait for the container to build and pip install to complete (~2 min)
  3. Copy the env file and fill in your Azure credentials:
 cp .env.example .env
 # edit .env with your Azure endpoint URLs and API keys

Microsoft Foundry — Endpoint URL Format

The correct base_url for each model uses the **/providers/cohere** path (visible on the deployment's Details page in the Microsoft Foundry portal under Target URI):

Model type Target URI pattern SDK base_url
Embed …/providers/cohere/v2/embed https://<resource>.services.ai.azure.com/providers/cohere
Rerank …/providers/cohere/v2/rerank https://<resource>.services.ai.azure.com/providers/cohere
Parse …/providers/cohere/v2/parse https://<resource>.services.ai.azure.com/providers/cohere

Note: cohere.ClientV2 automatically appends /v2/embed, /v2/rerank, or /v2/parse to the base_url.

Environment Variables

Variable Description
EMBED_MODEL Cohere embed model name (e.g. Embed-V5-Pro)
EMBED_BASE_URL Microsoft Foundry endpoint — up to /providers/cohere
EMBED_API_KEY Azure API key for the embed deployment
RERANK_MODEL Cohere rerank model name (e.g. Cohere-rerank-v4.0-pro)
RERANK_BASE_URL Microsoft Foundry endpoint — up to /providers/cohere
RERANK_API_KEY Azure API key for the rerank deployment
PARSE_MODEL Cohere parse model name (e.g. parse-v5.0)
PARSE_BASE_URL Microsoft Foundry endpoint — up to /providers/cohere
PARSE_API_KEY Azure API key for the parse deployment
CHROMA_DB_PATH Path to ChromaDB storage (default: ./chroma_db)