Merge branch 'main' into struct-op-tools

Signed-off-by: Meor Amer <[email protected]>

.github/workflows/check-mdx-frontmatter.yml

@@ -26,4 +26,4 @@ jobs:
         run: pnpm install
 
       - name: Run MDX frontmatter check
-        run: node .github/scripts/check-mdx-frontmatter.js
+        run: node .github/scripts/check-mdx-frontmatter.cjs

.github/workflows/snippet-ci.yml

@@ -0,0 +1,40 @@
+name: snippet-ci
+
+on:
+  pull_request: {}
+  push:
+    branches:
+      - main
+
+jobs:
+  run:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v2
+        with:
+          version: 8
+
+      - name: Install Dependencies
+        shell: bash
+        run: pnpm install             
+
+      - name: Set up python
+        uses: actions/setup-python@v2
+        with:
+          python-version: '3.x'
+
+      - name: poetry install
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install poetry
+          poetry install
+
+      - name: Run snippet tests
+        continue-on-error: true
+        env:
+          CO_API_KEY: ${{ secrets.COHERE_TOKEN }}
+        run: pnpm run --filter snippet-tester test

fern/docs.yml

@@ -571,6 +571,9 @@ redirects:
   - source: /docs/data-statement
     destination: /docs/usage-guidelines
     permanent: true
+  - source: /docs/usage-guidelines
+    destination: /docs/usage-policy
+    permanent: true
   - source: /v2/v2/:slug* 
     destination: /v2/:slug*
     permanent: true

fern/pages/cookbooks/convfinqa-finetuning-wandb.mdx

@@ -55,10 +55,10 @@ from cohere.finetuning import (
 )
 
 # fill in your Cohere API key here
-os.environ['COHERE_API_KEY'] = "<COHERE_API_KEY>" 
+os.environ["COHERE_API_KEY"] = "<COHERE_API_KEY>"
 
 # instantiate the Cohere client
-co = cohere.Client(os.environ['COHERE_API_KEY'])
+co = cohere.Client(os.environ["COHERE_API_KEY"])
 ```
 
 ## Dataset

fern/pages/cookbooks/deploy-finetuned-model-aws-marketplace.mdx

@@ -71,8 +71,8 @@ To subscribe to the algorithm:
 Install the Python packages you will use below and import them. For example, you can run the command below to install `cohere` if you haven't done so.
 
 
-```python
-!pip install "cohere>=5.11.0"
+```sh
+pip install "cohere>=5.11.0"
 ```
 
 
@@ -200,9 +200,10 @@ save_hf_model(merged_weights_dir, merged_model)
 
 
 ```python
-%%time
 sess = sage.Session()
-merged_weights = S3Uploader.upload(merged_weights_dir, s3_checkpoint_dir, sagemaker_session=sess)
+merged_weights = S3Uploader.upload(
+    merged_weights_dir, s3_checkpoint_dir, sagemaker_session=sess
+)
 print("merged_weights", merged_weights)
 ```
 
@@ -213,7 +214,6 @@ Create Cohere client and use it to export the merged weights to the TensorRT-LLM
 
 
 ```python
-%%time
 co = cohere.SagemakerClient(aws_region=region)
 co.sagemaker_finetuning.export_finetune(
     arn=arn,
@@ -232,7 +232,6 @@ The Cohere client provides a built-in method to create an endpoint for inference
 
 
 ```python
-%%time
 co.sagemaker_finetuning.create_endpoint(
     arn=arn,
     endpoint_name=endpoint_name,

fern/pages/cookbooks/finetune-on-sagemaker.mdx

@@ -58,10 +58,11 @@ To subscribe to the model algorithm:
 2. On the AWS Marketplace listing, click on the **Continue to Subscribe** button.
 3. On the **Subscribe to this software** page, review and click on **"Accept Offer"** if you and your organization agrees with EULA, pricing, and support terms. On the "Configure and launch" page, make sure ARN displayed in your region match with the ARN in the following cell.
 
+```sh
+pip install "cohere>=5.11.0"
+```
 
 ```python
-!pip install "cohere>=5.11.0"
-
 import cohere
 import boto3
 import sagemaker as sage
@@ -297,7 +298,7 @@ from tqdm import tqdm
 total = 0
 correct = 0
 for line in tqdm(
-    open('./sample_finetune_scienceQA_eval.jsonl').readlines()
+    open("./sample_finetune_scienceQA_eval.jsonl").readlines()
 ):
     total += 1
     question_answer_json = json.loads(line)

fern/pages/cookbooks/rag-cohere-mongodb.mdx

@@ -52,8 +52,8 @@ Libraries:
 
 
 
-```python
-!pip install --quiet datasets tqdm cohere pymongo
+```sh
+pip install --quiet datasets tqdm cohere pymongo
 ```
 
 
@@ -183,11 +183,11 @@ def combine_attributes(row):
     combined = f"{row['company']} {row['sector']} "
 
     # Add reports information
-    for report in row['reports']:
+    for report in row["reports"]:
         combined += f"{report['year']} {report['title']} {report['author']} {report['content']} "
 
     # Add recent news information
-    for news in row['recent_news']:
+    for news in row["recent_news"]:
         combined += f"{news['headline']} {news['summary']} "
 
     return combined.strip()
@@ -196,15 +196,15 @@ def combine_attributes(row):
 
 ```python
 # Add the new column 'combined_attributes'
-dataset_df['combined_attributes'] = dataset_df.apply(
+dataset_df["combined_attributes"] = dataset_df.apply(
     combine_attributes, axis=1
 )
 ```
 
 
 ```python
 # Display the first few rows of the updated dataframe
-dataset_df[['company', 'ticker', 'combined_attributes']].head()
+dataset_df[["company", "ticker", "combined_attributes"]].head()
 ```
 
 <div>
@@ -270,7 +270,7 @@ def get_embedding(
         texts=[text],
         model=model,
         input_type=input_type,  # Used for embeddings of search queries run against a vector DB to find relevant documents
-        embedding_types=['float'],
+        embedding_types=["float"],
     )
 
     return response.embeddings.float[0]
@@ -279,7 +279,7 @@ def get_embedding(
 # Apply the embedding function with a progress bar
 tqdm.pandas(desc="Generating embeddings")
 dataset_df["embedding"] = dataset_df[
-    'combined_attributes'
+    "combined_attributes"
 ].progress_apply(get_embedding)
 
 print(f"We just computed {len(dataset_df['embedding'])} embeddings.")
@@ -421,8 +421,8 @@ def get_mongo_client(mongo_uri):
     )
 
     # Validate the connection
-    ping_result = client.admin.command('ping')
-    if ping_result.get('ok') == 1.0:
+    ping_result = client.admin.command("ping")
+    if ping_result.get("ok") == 1.0:
         # Connection successful
         print("Connection to MongoDB successful")
         return client
@@ -478,7 +478,7 @@ MongoDB's Document model and its compatibility with Python dictionaries offer se
 ![](../../assets/images/rag-cohere-mongodb-4.png)
 
 ```python
-documents = dataset_df.to_dict('records')
+documents = dataset_df.to_dict("records")
 collection.insert_many(documents)
 
 print("Data ingestion into MongoDB completed")
@@ -592,13 +592,13 @@ def rerank_documents(query: str, documents, top_n: int = 3):
             original_doc = documents[result.index]
             top_documents_after_rerank.append(
                 {
-                    'company': original_doc['company'],
-                    'combined_attributes': original_doc[
-                        'combined_attributes'
+                    "company": original_doc["company"],
+                    "combined_attributes": original_doc[
+                        "combined_attributes"
                     ],
-                    'reports': original_doc['reports'],
-                    'vector_search_score': original_doc['score'],
-                    'relevance_score': result.relevance_score,
+                    "reports": original_doc["reports"],
+                    "vector_search_score": original_doc["score"],
+                    "relevance_score": result.relevance_score,
                 }
             )
 
@@ -724,9 +724,9 @@ pd.DataFrame(reranked_documents).head()
 def format_documents_for_chat(documents):
     return [
         {
-            "company": doc['company'],
+            "company": doc["company"],
             # "reports": doc['reports'],
-            "combined_attributes": doc['combined_attributes'],
+            "combined_attributes": doc["combined_attributes"],
         }
         for doc in documents
     ]
@@ -825,7 +825,7 @@ class CohereChat:
         # Use the connection string from history_params
         self.client = pymongo.MongoClient(
             self.history_params.get(
-                'connection_string', 'mongodb://localhost:27017/'
+                "connection_string", "mongodb://localhost:27017/"
             )
         )
 
@@ -838,34 +838,34 @@ class CohereChat:
         # Use the history_collection from history_params, or default to "chat_history"
         self.history_collection = self.db[
             self.history_params.get(
-                'history_collection', 'chat_history'
+                "history_collection", "chat_history"
             )
         ]
 
         # Use the session_id from history_params, or default to "default_session"
         self.session_id = self.history_params.get(
-            'session_id', 'default_session'
+            "session_id", "default_session"
         )
 
     def add_to_history(self, message: str, prefix: str = ""):
         self.history_collection.insert_one(
             {
-                'session_id': self.session_id,
-                'message': message,
-                'prefix': prefix,
+                "session_id": self.session_id,
+                "message": message,
+                "prefix": prefix,
             }
         )
 
     def get_chat_history(self) -> List[Dict[str, str]]:
         history = self.history_collection.find(
-            {'session_id': self.session_id}
-        ).sort('_id', 1)
+            {"session_id": self.session_id}
+        ).sort("_id", 1)
         return [
             {
                 "role": (
-                    "user" if item['prefix'] == "USER" else "chatbot"
+                    "user" if item["prefix"] == "USER" else "chatbot"
                 ),
-                "message": item['message'],
+                "message": item["message"],
             }
             for item in history
         ]
@@ -875,11 +875,11 @@ class CohereChat:
     ) -> List[Dict]:
         rerank_docs = [
             {
-                'company': doc['company'],
-                'combined_attributes': doc['combined_attributes'],
+                "company": doc["company"],
+                "combined_attributes": doc["combined_attributes"],
             }
             for doc in documents
-            if doc['combined_attributes'].strip()
+            if doc["combined_attributes"].strip()
         ]
 
         if not rerank_docs:
@@ -897,11 +897,11 @@ class CohereChat:
 
             top_documents_after_rerank = [
                 {
-                    'company': rerank_docs[result.index]['company'],
-                    'combined_attributes': rerank_docs[result.index][
-                        'combined_attributes'
+                    "company": rerank_docs[result.index]["company"],
+                    "combined_attributes": rerank_docs[result.index][
+                        "combined_attributes"
                     ],
-                    'relevance_score': result.relevance_score,
+                    "relevance_score": result.relevance_score,
                 }
                 for result in response.results
             ]
@@ -925,8 +925,8 @@ class CohereChat:
     ) -> List[Dict]:
         return [
             {
-                "company": doc['company'],
-                "combined_attributes": doc['combined_attributes'],
+                "company": doc["company"],
+                "combined_attributes": doc["combined_attributes"],
             }
             for doc in documents
         ]
@@ -972,8 +972,8 @@ class CohereChat:
 
     def show_history(self):
         history = self.history_collection.find(
-            {'session_id': self.session_id}
-        ).sort('_id', 1)
+            {"session_id": self.session_id}
+        ).sort("_id", 1)
         for item in history:
             print(f"{item['prefix']}: {item['message']}")
             print("-------------------------")
@@ -988,9 +988,9 @@ chat = CohereChat(
     database=DB_NAME,
     main_collection=COLLECTION_NAME,
     history_params={
-        'connection_string': MONGO_URI,
-        'history_collection': "chat_history",
-        'session_id': 2,
+        "connection_string": MONGO_URI,
+        "history_collection": "chat_history",
+        "session_id": 2,
     },
 )
 

fern/pages/deployment-options/cohere-on-aws/amazon-sagemaker-setup-guide.mdx

@@ -115,6 +115,12 @@ To access Cohere's models on SageMaker Jumpstart, follow these steps:
 
 If you have any questions about this process, reach out to [email protected].
 
+## Optimize your Inference Latencies
+
+By default, SageMaker endpoints have a random routing strategy. This means that requests coming to the model endpoints are forwarded to the machine learning instances randomly, which can cause latency issues in applications focused on generative AI. In 2023, the SageMaker platform introduced a `RoutingStrategy` parameter allowing you to use the ‘least outstanding requests’ (LOR) approach to routing. With LOR, SageMaker monitors the load of the instances behind your endpoint as well as the models or inference components that are deployed on each instance, then optimally routes requests to the instance that is best suited to serve it.
+
+LOR has shown an improvement in latency under various conditions, and you can find more details [here](https://aws.amazon.com/blogs/machine-learning/minimize-real-time-inference-latency-by-using-amazon-sagemaker-routing-strategies/).
+
 ## Next Steps
 
 With your selected configuration and Product ARN available, you now have everything you need to integrate with Cohere’s model offerings on SageMaker.