Merge pull request #1 from Azure-Samples/githubmodels

Add support for GitHub models
Azure-Samples · Nov 27, 2024 · 80b1ee5 · 80b1ee5
2 parents b59400b + e1ed4d0
commit 80b1ee5
Show file tree

Hide file tree

Showing 15 changed files with 280 additions and 82 deletions.
diff --git a/.devcontainer/devcontainer.json b/.devcontainer/devcontainer.json
@@ -8,10 +8,11 @@
         "vscode": {
             "extensions": [
                 "ms-python.python",
-				"ms-python.vscode-pylance",
-				"charliermarsh.ruff",
+                "ms-python.vscode-pylance",
+                "charliermarsh.ruff",
                 "ms-azuretools.azure-dev",
-                "ms-azuretools.vscode-bicep"
+                "ms-azuretools.vscode-bicep",
+                "humao.rest-client"
             ],
             "python.defaultInterpreterPath": "/usr/local/bin/python",
             "[python]": {

diff --git a/.env.sample b/.env.sample
@@ -1,2 +1,5 @@
+OPENAI_HOST=azure # Options are azure or github
 AZURE_OPENAI_GPT_DEPLOYMENT=
 AZURE_OPENAI_SERVICE=
+AZURE_TENANT_ID=
+GITHUB_TOKEN=
diff --git a/README.md b/README.md
@@ -122,15 +122,36 @@ A related option is VS Code Dev Containers, which will open the project in your
 
 ## Running the Python examples
 
-1. If you're not already running in a Codespace or Dev Container, create a Python virtual environment.
+To run the samples, you'll either need to have already [deployed the Azure OpenAI account](#deployment) or use GitHub models.
 
-2. Install the requirements:
+1. Check that the `.env` file exists in the root of the project. If you [deployed an Azure OpenAI account](#deployment), it should have been created for you, and look like this:
+
+    ```shell
+    OPENAI_HOST=azure
+    AZURE_OPENAI_GPT_DEPLOYMENT=gpt-4o
+    AZURE_OPENAI_SERVICE=your-service-name
+    AZURE_TENANT_ID=your-tenant-id-1234
+    ```
+
+    If you're using GitHub models, create a `.env` file with the following content:
+
+    ```shell
+    OPENAI_HOST=github
+    GITHUB_TOKEN=
+    ```
+
+    You can create a GitHub token by following the [GitHub documentation](https://docs.github.com/en/github/authenticating-to-github/keeping-your-account-and-data-secure/creating-a-personal-access-token),
+    or open this project inside GitHub Codespaces where the token is already exposed as an environment variable.
+
+2. If you're not already running in a Codespace or Dev Container, create a Python virtual environment.
+
+3. Install the requirements:
 
     ```shell
     python -m pip install -r requirements.txt
     ```
 
-3. Run an example by running either `python example_file.py` or selecting the `Run` button on the opened file. Available examples:
+4. Run an example by running either `python example_file.py` or selecting the `Run` button on the opened file. Available examples:
 
     | Script filename       | Description                                                                 |
     |---------------------------|-----------------------------------------------------------------------------|

diff --git a/extract_github_issue.py b/extract_github_issue.py
@@ -12,18 +12,29 @@
 logging.basicConfig(level=logging.WARNING)
 load_dotenv()
 
-# Configure Azure OpenAI
-if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
-    logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT environment variables are empty. See README.")
-    exit(1)
-credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
-token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
-client = openai.AzureOpenAI(
-    api_version="2024-08-01-preview",
-    azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
-    azure_ad_token_provider=token_provider,
-)
-model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+
+if os.getenv("OPENAI_HOST", "azure") == "azure":
+    if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
+        logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT env variables are empty. See README.")
+        exit(1)
+    credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
+    token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
+    client = openai.AzureOpenAI(
+        api_version="2024-08-01-preview",
+        azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
+        azure_ad_token_provider=token_provider,
+    )
+    model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+else:
+    if not os.getenv("GITHUB_TOKEN"):
+        logging.warning("GITHUB_TOKEN env variable is empty. See README.")
+        exit(1)
+    client = openai.OpenAI(
+        base_url="https://models.inference.ai.azure.com",
+        api_key=os.environ["GITHUB_TOKEN"],
+        # Specify the API version to use the Structured Outputs feature
+        default_query={"api-version": "2024-08-01-preview"})
+    model_name = "gpt-4o"
 
 
 # Define models for Structured Outputs

diff --git a/extract_github_repo.py b/extract_github_repo.py
@@ -13,17 +13,28 @@
 logging.basicConfig(level=logging.WARNING)
 load_dotenv()
 
-# Configure Azure OpenAI
-if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
-    logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT environment variables are empty. See README.")
-    exit(1)
-credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
-token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
-client = openai.AzureOpenAI(
-    api_version="2024-08-01-preview",
-    azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
-    azure_ad_token_provider=token_provider,
-)
+if os.getenv("OPENAI_HOST", "azure") == "azure":
+    if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
+        logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT env variables are empty. See README.")
+        exit(1)
+    credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
+    token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
+    client = openai.AzureOpenAI(
+        api_version="2024-08-01-preview",
+        azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
+        azure_ad_token_provider=token_provider,
+    )
+    model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+else:
+    if not os.getenv("GITHUB_TOKEN"):
+        logging.warning("GITHUB_TOKEN env variable is empty. See README.")
+        exit(1)
+    client = openai.OpenAI(
+        base_url="https://models.inference.ai.azure.com",
+        api_key=os.environ["GITHUB_TOKEN"],
+        # Specify the API version to use the Structured Outputs feature
+        default_query={"api-version": "2024-08-01-preview"})
+    model_name = "gpt-4o"
 
 
 # Define models for Structured Outputs
@@ -69,7 +80,7 @@ class RepoOverview(BaseModel):
 
 # Send request to GPT model to extract using Structured Outputs
 completion = client.beta.chat.completions.parse(
-    model=os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"),
+    model=model_name,
     messages=[
         {
             "role": "system",

diff --git a/extract_image_graph.py b/extract_image_graph.py
@@ -11,18 +11,28 @@
 logging.basicConfig(level=logging.WARNING)
 load_dotenv()
 
-# Configure Azure OpenAI
-if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
-    logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT environment variables are empty. See README.")
-    exit(1)
-credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
-token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
-client = openai.AzureOpenAI(
-    api_version="2024-08-01-preview",
-    azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
-    azure_ad_token_provider=token_provider,
-)
-model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+if os.getenv("OPENAI_HOST", "azure") == "azure":
+    if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
+        logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT env variables are empty. See README.")
+        exit(1)
+    credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
+    token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
+    client = openai.AzureOpenAI(
+        api_version="2024-08-01-preview",
+        azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
+        azure_ad_token_provider=token_provider,
+    )
+    model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+else:
+    if not os.getenv("GITHUB_TOKEN"):
+        logging.warning("GITHUB_TOKEN env variable is empty. See README.")
+        exit(1)
+    client = openai.OpenAI(
+        base_url="https://models.inference.ai.azure.com",
+        api_key=os.environ["GITHUB_TOKEN"],
+        # Specify the API version to use the Structured Outputs feature
+        default_query={"api-version": "2024-08-01-preview"})
+    model_name = "gpt-4o"
 
 
 # Define models for Structured Outputs
@@ -46,7 +56,7 @@ def open_image_as_base64(filename):
 
 # Send request to GPT model to extract using Structured Outputs
 completion = client.beta.chat.completions.parse(
-    model=os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"),
+    model=model_name,
     messages=[
         {"role": "system", "content": "Extract the information from the graph"},
         {

diff --git a/extract_image_table.py b/extract_image_table.py
@@ -11,17 +11,28 @@
 logging.basicConfig(level=logging.WARNING)
 load_dotenv()
 
-# Configure Azure OpenAI
-if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
-    logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT environment variables are empty. See README.")
-    exit(1)
-credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
-token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
-client = openai.AzureOpenAI(
-    api_version="2024-08-01-preview",
-    azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
-    azure_ad_token_provider=token_provider,
-)
+if os.getenv("OPENAI_HOST", "azure") == "azure":
+    if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
+        logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT env variables are empty. See README.")
+        exit(1)
+    credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
+    token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
+    client = openai.AzureOpenAI(
+        api_version="2024-08-01-preview",
+        azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
+        azure_ad_token_provider=token_provider,
+    )
+    model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+else:
+    if not os.getenv("GITHUB_TOKEN"):
+        logging.warning("GITHUB_TOKEN env variable is empty. See README.")
+        exit(1)
+    client = openai.OpenAI(
+        base_url="https://models.inference.ai.azure.com",
+        api_key=os.environ["GITHUB_TOKEN"],
+        # Specify the API version to use the Structured Outputs feature
+        default_query={"api-version": "2024-08-01-preview"})
+    model_name = "gpt-4o"
 
 
 # Define models for Structured Outputs
@@ -53,7 +64,7 @@ def open_image_as_base64(filename):
 
 # Send request to GPT model to extract using Structured Outputs
 completion = client.beta.chat.completions.parse(
-    model=os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"),
+    model=model_name,
     messages=[
         {"role": "system", "content": "Extract the information from the table"},
         {

diff --git a/extract_pdf_receipt.py b/extract_pdf_receipt.py
@@ -11,18 +11,28 @@
 logging.basicConfig(level=logging.WARNING)
 load_dotenv()
 
-# Configure Azure OpenAI
-if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
-    logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT environment variables are empty. See README.")
-    exit(1)
-credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
-token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
-client = openai.AzureOpenAI(
-    api_version="2024-08-01-preview",
-    azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
-    azure_ad_token_provider=token_provider,
-)
-model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+if os.getenv("OPENAI_HOST", "azure") == "azure":
+    if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
+        logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT env variables are empty. See README.")
+        exit(1)
+    credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
+    token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
+    client = openai.AzureOpenAI(
+        api_version="2024-08-01-preview",
+        azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
+        azure_ad_token_provider=token_provider,
+    )
+    model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+else:
+    if not os.getenv("GITHUB_TOKEN"):
+        logging.warning("GITHUB_TOKEN env variable is empty. See README.")
+        exit(1)
+    client = openai.OpenAI(
+        base_url="https://models.inference.ai.azure.com",
+        api_key=os.environ["GITHUB_TOKEN"],
+        # Specify the API version to use the Structured Outputs feature
+        default_query={"api-version": "2024-08-01-preview"})
+    model_name = "gpt-4o"
 
 
 # Define models for Structured Outputs
@@ -45,7 +55,7 @@ class Receipt(BaseModel):
 
 # Send request to GPT model to extract using Structured Outputs
 completion = client.beta.chat.completions.parse(
-    model=os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"),
+    model=model_name,
     messages=[
         {"role": "system", "content": "Extract the information from the blog post"},
         {"role": "user", "content": md_text},

diff --git a/extract_webpage.py b/extract_webpage.py
@@ -12,19 +12,28 @@
 logging.basicConfig(level=logging.WARNING)
 load_dotenv()
 
-# Configure Azure OpenAI
-if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
-    logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT environment variables are empty. See README.")
-    exit(1)
-credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
-token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
-client = openai.AzureOpenAI(
-    api_version="2024-08-01-preview",
-    azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
-    azure_ad_token_provider=token_provider,
-)
-model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
-
+if os.getenv("OPENAI_HOST", "azure") == "azure":
+    if not os.getenv("AZURE_OPENAI_SERVICE") or not os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"):
+        logging.warning("AZURE_OPENAI_SERVICE and AZURE_OPENAI_GPT_DEPLOYMENT env variables are empty. See README.")
+        exit(1)
+    credential = azure.identity.AzureDeveloperCliCredential(tenant_id=os.getenv("AZURE_TENANT_ID"))
+    token_provider = azure.identity.get_bearer_token_provider(credential, "https://cognitiveservices.azure.com/.default")
+    client = openai.AzureOpenAI(
+        api_version="2024-08-01-preview",
+        azure_endpoint=f"https://{os.getenv('AZURE_OPENAI_SERVICE')}.openai.azure.com",
+        azure_ad_token_provider=token_provider,
+    )
+    model_name = os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT")
+else:
+    if not os.getenv("GITHUB_TOKEN"):
+        logging.warning("GITHUB_TOKEN env variable is empty. See README.")
+        exit(1)
+    client = openai.OpenAI(
+        base_url="https://models.inference.ai.azure.com",
+        api_key=os.environ["GITHUB_TOKEN"],
+        # Specify the API version to use the Structured Outputs feature
+        default_query={"api-version": "2024-08-01-preview"})
+    model_name = "gpt-4o"
 
 # Define models for Structured Outputs
 class BlogPost(BaseModel):
@@ -46,7 +55,7 @@ class BlogPost(BaseModel):
 
 # Send request to GPT model to extract using Structured Outputs
 completion = client.beta.chat.completions.parse(
-    model=os.getenv("AZURE_OPENAI_GPT_DEPLOYMENT"),
+    model=model_name,
     messages=[
         {"role": "system", "content": "Extract the information from the blog post"},
         {"role": "user", "content": f"{post_title}\n{post_contents}"},

diff --git a/http/.env.azure b/http/.env.azure
@@ -0,0 +1,3 @@
+AZURE_OPENAI_SERVICE=
+AZURE_OPENAI_GPT_DEPLOYMENT=
+AZURE_OPENAI_TOKEN=
diff --git a/http/.env.github b/http/.env.github
@@ -0,0 +1,2 @@
+# In Codespaces, you can find this value in the env variables. Otherwise, generate a new Personal Access Token in GitHub.
+GITHUB_TOKEN=
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@
		# In Codespaces, you can find this value in the env variables. Otherwise, generate a new Personal Access Token in GitHub.
		GITHUB_TOKEN=