Integrations

docs/embeddings.md

@@ -23,6 +23,8 @@ Chroma provides lightweight wrappers around popular embedding providers, making
 
 We welcome pull requests to add new Embedding Functions to the community.
 
+[➕ Add New](/embeddings/add)
+
 ***
 
 ## Default: all-MiniLM-L6-v2

docs/embeddings/add.md

@@ -0,0 +1,41 @@
+---
+title: ➕ Add
+---
+
+New embeddings integrations help Chroma users build better applications, faster! 
+
+### Process
+
+To land a new embedding integration, you will want to add the integrations to the Python and JS clients. 
+- [Python embedding functions](https://github.com/chroma-core/chroma/blob/12785d71ea476ef3cbd28b419e7807bf7f2129d3/chromadb/utils/embedding_functions.py#L4)
+- [JS embedding functions](https://github.com/chroma-core/chroma/tree/main/clients/js/src/embeddings)
+
+We will accept embedding functions with just one or the other in some circumstances, but we really like to keep parity between them.
+
+The process is:
+1. Please open a new PR in the [Chroma Repo](https://github.com/chroma-core/chroma). 
+2. Please open a new PR in the [Docs Repo](https://github.com/chroma-core/docs).
+3. Chroma will approve and land the new embedding functions and cut a new Python/JS release
+4. The docs PR will merge and be deployed
+
+### Embedding Page
+
+We strongly encourage having the following sections:
+- Overview
+- Info table on available models and their details
+- Simple getting-started usage for Python and JS
+- Advanced usage section with examples and links to other resources
+
+The [OpenAI](/embeddings/openai) integration is a great reference! 
+
+## Integration Overview and Sidebar
+
+You will also want to add your embedding functions to the embeddings overview page and sidebar.
+
+### Overview page
+
+Add your embedding function to the list here: https://github.com/chroma-core/docs/blob/main/docs/embeddings/index.md
+
+### Sidebar
+
+Add your embedding function to the sidebar: https://github.com/chroma-core/docs/blob/main/sidebars.js

docs/embeddings/cohere.md

@@ -3,56 +3,88 @@
 
 # Cohere
 
-import Tabs from '@theme/Tabs';
-import TabItem from '@theme/TabItem';
+Chroma also provides a convenient wrapper around Cohere's embedding API. This embedding function runs remotely on Cohere’s servers, and requires an API key. You can get an API key by signing up for an account at [Cohere](https://dashboard.cohere.ai/welcome/register).
 
-<div class="select-language">Select a language</div>
+<div class="data_table"></div>
 
-<Tabs queryString groupId="lang">
-<TabItem value="py" label="Python"></TabItem>
-<TabItem value="js" label="JavaScript"></TabItem>
-</Tabs>
+| Models | Input | Dimensionality | Context Size| 
+|--|--|--|--|--|
+|`embed-english-v3.0` | English | 1024 | 512 (recommended) | 
+|`embed-multilingual-v3.0` | [Full List](https://docs.cohere.com/docs/supported-languages) | 1024 | 512 (recommended) | 
+|`embed-english-light-v3.0` | English | 384 | 512 (recommended) | 
+|`embed-multilingual-light-v3.0` | [Full List](https://docs.cohere.com/docs/supported-languages) | 384 | 512 (recommended) | 
+|`embed-english-v2.0` | English | 4096 | 512 (recommended) | 
+|`embed-english-light-v2.0` | English | 1024 | 512 (recommended) |
+|`embed-multilingual-v2.0` | [Full List](https://docs.cohere.com/docs/supported-languages) | 768 | 512 (recommended) | 
 
-Chroma also provides a convenient wrapper around Cohere's embedding API. This embedding function runs remotely on Cohere’s servers, and requires an API key. You can get an API key by signing up for an account at [Cohere](https://dashboard.cohere.ai/welcome/register).
 
-<Tabs queryString groupId="lang" className="hideTabSwitcher">
-<TabItem value="py" label="Python">
+## Basic Usage
 
-This embedding function relies on the `cohere` python package, which you can install with `pip install cohere`.
+### Python
+
+```bash
+pip install cohere
+```
 
 ```python
-cohere_ef  = embedding_functions.CohereEmbeddingFunction(api_key="YOUR_API_KEY",  model_name="large")
-cohere_ef(texts=["document1","document2"])
+
+from chromadb.utils import embedding_functions
+
+embedder = embedding_functions.CohereEmbeddingFunction(
+        api_key="YOUR_API_KEY")
+
+collection = client.create_collection(
+        name="cohere_ef", 
+        embedding_function=embedder)
 ```
 
-</TabItem>
-<TabItem value="js" label="JavaScript">
+### Javascript
+
+```bash
+yarn add cohere-ai
+```
 
 ```javascript
-const {CohereEmbeddingFunction} = require('chromadb');
-const embedder = new CohereEmbeddingFunction("apiKey")
+import { ChromaClient, CohereEmbeddingFunction } from 'chromadb'
 
-// use directly 
-const embeddings = embedder.generate(["document1","document2"])
+const embedder = new CohereEmbeddingFunction({
+    apiKey: "YOUR_API_KEY"
+})
 
-// pass documents to query for .add and .query
-const collection = await client.createCollection({name: "name", embeddingFunction: embedder})
-const collectionGet = await client.getCollection({name:"name", embeddingFunction: embedder})
+const collection = await client.createCollection({
+    name: "cohere_ef", 
+    embeddingFunction: embedder
+})
 ```
 
-</TabItem>
+## Advanced Usage
 
-</Tabs>
+### Call directly 
 
+By passing the embedding function to a Collection, Chroma handles the embedding of documents and queries for you. However in some cases you may want to generate the embeddings outside and handle them yourself.
 
+#### Python
+
+```python
+embeddings = embedder(["document1","document2"])
+# [[0.04565250128507614, 0.01611952856183052...], [0.030171213671565056, 0.007690359838306904...]]
+```
+
+#### Javascript
+
+```javascript
+const embeddings = embedder.generate(["document1","document2"])
+// [[0.04565250128507614, 0.01611952856183052...], [0.030171213671565056, 0.007690359838306904...]]
+```
+
+### Using a different model
 
 You can pass in an optional `model_name` argument, which lets you choose which Cohere embeddings model to use. By default, Chroma uses `large` model. You can see the available models under `Get embeddings` section [here](https://docs.cohere.ai/reference/embed).
 
 
 ### Multilingual model example
 
-<Tabs queryString groupId="lang" className="hideTabSwitcher">
-<TabItem value="py" label="Python">
+#### Python
 
 ```python
 cohere_ef  = embedding_functions.CohereEmbeddingFunction(
@@ -69,11 +101,10 @@ cohere_ef(texts=multilingual_texts)
 
 ```
 
-</TabItem>
-<TabItem value="js" label="JavaScript">
+#### Javascript
 
 ```javascript
-const {CohereEmbeddingFunction} = require('chromadb');
+import { CohereEmbeddingFunction } from 'chromadb'
 const embedder = new CohereEmbeddingFunction("apiKey")
 
 multilingual_texts  = [ 'Hello from Cohere!', 'مرحبًا من كوهير!', 
@@ -86,11 +117,4 @@ const embeddings = embedder.generate(multilingual_texts)
 
 ```
 
-
-</TabItem>
-
-</Tabs>
-
-
-
 For more information on multilingual model you can read [here](https://docs.cohere.ai/docs/multilingual-language-models).

docs/embeddings/google-gemini.md

@@ -3,25 +3,24 @@
 
 # Google Generative AI
 
-import Tabs from '@theme/Tabs';
-import TabItem from '@theme/TabItem';
+Chroma also provides a convenient wrapper around Google's embedding API. This embedding function runs remotely on Google servers, and requires an API key. You can get an API key by signing up for an account at [Google MakerSuite](https://makersuite.google.com/).
 
-<div class="select-language">Select a language</div>
+<div class="data_table"></div>
 
-<Tabs queryString groupId="lang">
-<TabItem value="py" label="Python"></TabItem>
-<TabItem value="js" label="JavaScript"></TabItem>
-</Tabs>
+| Models | Input | Dimensionality | 
+|--|--|--|--|--|
+|`models/embedding-001` | English | 768 | 
 
-Chroma provides a convenient wrapper around Google's Generative AI embedding API. This embedding function runs remotely on Google's servers, and requires an API key. 
+## Basic Usage
 
-You can get an API key by signing up for an account at [Google MakerSuite](https://makersuite.google.com/).
-
-<Tabs queryString groupId="lang" className="hideTabSwitcher">
-<TabItem value="py" label="Python">
+### Python
 
 This embedding function relies on the `google-generativeai` python package, which you can install with `pip install google-generativeai`.
 
+```bash
+pip install google-generativeai
+```
+
 ```python
 # import
 import chromadb
@@ -40,11 +39,14 @@ You can view a more [complete example](https://github.com/chroma-core/chroma/tre
 
 For more info - please visit the [official Google python docs](https://ai.google.dev/tutorials/python_quickstart).
 
-</TabItem>
-<TabItem value="js" label="JavaScript">
+### Javascript
 
 This embedding function relies on the `@google/generative-ai` npm package, which you can install with `yarn add @google/generative-ai`.
 
+```bash
+yarn add @google/generative-ai
+```
+
 ```javascript
 import { ChromaClient, GoogleGenerativeAiEmbeddingFunction } from 'chromadb'
 const embedder = new GoogleGenerativeAiEmbeddingFunction({googleApiKey: "<YOUR API KEY>"})
@@ -61,7 +63,62 @@ You can view a more [complete example using Node](https://github.com/chroma-core
 
 For more info - please visit the [official Google JS docs](https://ai.google.dev/tutorials/node_quickstart).
 
-</TabItem>
+## Advanced Usage
+
+### Call directly 
+
+By passing the embedding function to a Collection, Chroma handles the embedding of documents and queries for you. However in some cases you may want to generate the embeddings outside and handle them yourself.
+
+#### Python
+
+```python
+embeddings = embedder(["document1","document2"])
+# [[0.04565250128507614, 0.01611952856183052...], [0.030171213671565056, 0.007690359838306904...]]
+```
+
+#### Javascript
+
+```javascript
+const embeddings = embedder.generate(["document1","document2"])
+// [[0.04565250128507614, 0.01611952856183052...], [0.030171213671565056, 0.007690359838306904...]]
+```
+
+
+
+### Task Type
+
+Google's Embedding endpoint also accepts a `task_type`/`taskType` parameter. This may boost performance for your specific usage. 
+
+<div class="data_table"></div>
+
+| Task Type|	Description|
+|--|--|
+|RETRIEVAL_QUERY |	Specifies the given text is a query in a search/retrieval setting.|
+|RETRIEVAL_DOCUMENT	|Specifies the given text is a document in a search/retrieval setting. Using this task type requires a title.|
+|SEMANTIC_SIMILARITY|	Specifies the given text will be used for Semantic Textual Similarity (STS).|
+|CLASSIFICATION	|Specifies that the embeddings will be used for classification.|
+|CLUSTERING	|Specifies that the embeddings will be used for clustering.|
+
+Here is a python demonstration of how to use `RETRIEVAL_QUERY` with `RETRIEVAL_DOCUMENT`. 
+
+```python
+# import
+import chromadb
+from chromadb.utils import embedding_functions
+
+google_ef  = embedding_functions.GoogleGenerativeAiEmbeddingFunction(api_key="YOUR_API_KEY", task_type='RETRIEVAL_DOCUMENT')
+
+# pass documents to query for .add and .query
+collection = client.create_collection(name="name", embedding_function=google_ef)
+
+# add your documents
+collection.add(...) 
+
+# create a new EF for Query and re-get your collection
+google_ef2 = embedding_functions.GoogleGenerativeAiEmbeddingFunction(api_key="YOUR_API_KEY", task_type='RETRIEVAL_QUERY')
+collection = client.get_collection(name="name", embedding_function=google_ef2)
 
-</Tabs>
+# query your documents
+collection.query(...) 
 
+```