klas/t6_mem0
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add support: Custom instruction/categories for projects (#2134)

Browse Source
This commit is contained in:
Dev Khant
2025-01-11 13:38:20 +05:30
committed by GitHub
parent ac8cf59473
commit 7f0d766c09
7 changed files with 659 additions and 192 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/api-reference/project/get-instructions-and-categories.mdx Normal file
View File

@@ -0,0 +1,4 @@
---
title: 'Get Custom Instructions and Categories'
openapi: get /api/v1/orgs/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/
---

6
docs/api-reference/project/update-instructions-and-categories.mdx Normal file
View File

@@ -0,0 +1,6 @@
---
title: 'Update Custom Instructions and Categories'
openapi: post /api/v1/orgs/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/
---
Please refer to the [how to use custom instructions/categories](/features/custom-instructions) for more information.

50
docs/features/custom-categories.mdx
View File

@@ -45,7 +45,55 @@ Mostly cook at home because of gym plan. (fitness, cooking)
<Note> The more detailed the description of categories is, the better output the user will receive. When custom categories are provided in the `add` API call, they will completely replace the default categories and will be directly assigned to the memory, so make sure to include all categories you want to use. </Note>
<Note> We will soon release a feature that allows users to set custom categories in `project`. This will allow the functionality where relevant categories are automatically assigned to the memory based on the input text provided. </Note>
## Setting Project-Level Custom Categories
You can also set custom categories at the project level, which will be applied to all memories added within that project. We will automatically assign relevant categories from your custom set to new memories based on their content.
Here's how to set custom categories:
<CodeGroup>
```python Code
# Update custom categories
new_categories = [
{"cooking": "For users interested in cooking and culinary experiences"},
{"fitness": "Content related to fitness and exercise"}
]
response = client.update_custom_instructions_and_categories({"custom_categories": new_categories})
print(response)
```
```json Output
{
"message": "Updated custom categories"
}
```
</CodeGroup>
You can also retrieve the current custom categories:
<CodeGroup>
```python Code
from mem0 import MemoryClient
client = MemoryClient(api_key="xxx")
# Get current custom categories
categories = client.get_custom_instructions_and_categories(["custom_categories"])
print(categories)
```
```json Output
{
"custom_categories": [
{"cooking": "For users interested in cooking and culinary experiences"},
{"fitness": "Content related to fitness and exercise"}
]
}
```
</CodeGroup>
These project-level categories will be automatically applied to all new memories added to the project.
## Default Categories
Here is the list of **default categories**. Ensure you review these before creating custom categories to prevent duplication.

59
docs/features/custom-instructions.mdx Normal file
View File

@@ -0,0 +1,59 @@
---
title: Custom Instructions
description: 'Enhance your product experience by adding custom instructions tailored to your needs'
---
## Introduction to Custom Instructions
Custom instructions allow you to define specific guidelines for your project. This feature helps ensure consistency and provides clear direction for handling project-specific requirements.
Custom instructions are particularly useful when you want to:
- Define how information should be extracted from conversations
- Specify what types of data should be captured or ignored
- Set rules for categorizing and organizing memories
- Maintain consistent handling of project-specific requirements
When custom instructions are set at the project level, they will be applied to all new memories added within that project. This ensures that your data is processed according to your defined guidelines across your entire project.
## Setting Custom Instructions
You can set custom instructions for your project using the following method:
<CodeGroup>
```python Code
# Update custom instructions
prompt = """
Your task is to extract ONLY health-related information from conversations, including:
- Medical conditions, symptoms, diagnoses
- Medications, treatments, procedures
- Diet, exercise, sleep habits
- Doctor visits and appointments
- Health metrics (weight, blood pressure, etc)
"""
response = client.update_custom_instructions_and_categories({"custom_instructions": prompt})
print(response)
```
```json Output
{
"message": "Updated custom instructions"
}
```
</CodeGroup>
You can also retrieve the current custom instructions:
<CodeGroup>
```python Code
# Retrieve current custom instructions
response = client.get_custom_instructions_and_categories(["custom_instructions"])
print(response)
```
```json Output
{
"custom_instructions": "Your task is to extract ONLY health-related information from conversations, including:\n\n- Medical conditions, symptoms, diagnoses\n- Medications, treatments, procedures\n- Diet, exercise, sleep habits\n- Doctor visits and appointments\n- Health metrics (weight, blood pressure, etc)"
}
```
</CodeGroup>

4
docs/mint.json
View File

@@ -64,7 +64,7 @@
"platform/quickstart",
{
"group": "Features",
"pages": ["features/selective-memory", "features/custom-categories", "features/direct-import", "features/async-client", "features/memory-export"]
"pages": ["features/selective-memory", "features/custom-categories", "features/custom-instructions", "features/direct-import", "features/async-client", "features/memory-export"]
},
"features/langchain-tools"
]
@@ -197,6 +197,8 @@
"api-reference/project/get-project",
"api-reference/project/create-project",
"api-reference/project/delete-project",
"api-reference/project/get-instructions-and-categories",
"api-reference/project/update-instructions-and-categories",
{
"group": "Members APIs",
"pages":[

191
docs/openapi.json
View File

@@ -4044,6 +4044,197 @@
}
]
}
},
"/api/v1/orgs/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/": {
"get": {
"tags": ["custom-instructions-and-categories"],
"summary": "Get custom instructions and categories",
"description": "Retrieve custom categories and instructions for a specific project",
"operationId": "get_custom_instructions_and_categories",
"parameters": [
{
"name": "org_id",
"in": "path",
"required": true,
"schema": {
"type": "string"
},
"description": "Organization ID"
},
{
"name": "project_id",
"in": "path",
"required": true,
"schema": {
"type": "string"
},
"description": "Project ID"
}
],
"responses": {
"200": {
"description": "Successfully retrieved custom instructions and categories",
"content": {
"application/json": {
"schema": {
"type": "object",
"properties": {
"custom_categories": {
"type": "array",
"items": {
"type": "string"
},
"description": "List of custom categories for the project"
},
"custom_instructions": {
"type": "string",
"description": "Custom instructions text for the project"
}
},
"required": ["custom_categories", "custom_instructions"]
}
}
}
},
"403": {
"description": "Unauthorized access",
"content": {
"application/json": {
"schema": {
"type": "object",
"properties": {
"error": {
"type": "string",
"example": "Unauthorized to modify custom instructions and categories."
}
}
}
}
}
}
},
"x-code-samples": [
{
"lang": "Python",
"source": "import requests\n\nurl = \"https://api.mem0.ai/api/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/\"\nheaders = {\"Authorization\": \"Token <api-key>\"}\n\nresponse = requests.get(url, headers=headers)\nprint(response.json())"
},
{
"lang": "JavaScript",
"source": "fetch('https://api.mem0.ai/api/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/', {\n headers: {\n 'Authorization': 'Token <api-key>'\n }\n})\n.then(response => response.json())\n.then(data => console.log(data))\n.catch(error => console.error(error));"
}
]
},
"post": {
"tags": ["custom-instructions-and-categories"],
"summary": "Update custom instructions and categories",
"description": "Update custom categories and instructions for a specific project",
"operationId": "update_custom_instructions_and_categories",
"parameters": [
{
"name": "org_id",
"in": "path",
"required": true,
"schema": {
"type": "string"
},
"description": "Organization ID"
},
{
"name": "project_id",
"in": "path",
"required": true,
"schema": {
"type": "string"
},
"description": "Project ID"
}
],
"requestBody": {
"content": {
"application/json": {
"schema": {
"type": "object",
"properties": {
"custom_categories": {
"type": "array",
"items": {
"type": "string"
},
"description": "List of new custom categories for the project"
},
"custom_instructions": {
"type": "string",
"description": "New custom instructions text for the project"
}
}
}
}
}
},
"responses": {
"200": {
"description": "Successfully updated custom instructions and categories",
"content": {
"application/json": {
"schema": {
"type": "object",
"properties": {
"message": {
"type": "string",
"example": "Updated custom categories / instructions"
}
}
}
}
}
},
"400": {
"description": "Bad request",
"content": {
"application/json": {
"schema": {
"type": "object",
"additionalProperties": {
"type": "array",
"items": {
"type": "string"
}
},
"example": {
"field_name": ["error message"]
}
}
}
}
},
"403": {
"description": "Unauthorized access",
"content": {
"application/json": {
"schema": {
"type": "object",
"properties": {
"error": {
"type": "string",
"example": "Unauthorized to modify custom instructions and categories."
}
}
}
}
}
}
},
"x-code-samples": [
{
"lang": "Python",
"source": "import requests\n\nurl = \"https://api.mem0.ai/api/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/\"\nheaders = {\n \"Authorization\": \"Token <api-key>\",\n \"Content-Type\": \"application/json\"\n}\npayload = {\n \"custom_categories\": [\"new_category1\", \"new_category2\"],\n \"custom_instructions\": \"New custom instructions text\"\n}\n\nresponse = requests.post(url, headers=headers, json=payload)\nprint(response.json())"
},
{
"lang": "JavaScript",
"source": "fetch('https://api.mem0.ai/api/organizations/{org_id}/projects/{project_id}/custom-instructions-and-categories/', {\n method: 'POST',\n headers: {\n 'Authorization': 'Token <api-key>',\n 'Content-Type': 'application/json'\n },\n body: JSON.stringify({\n custom_categories: ['new_category1', 'new_category2'],\n custom_instructions: 'New custom instructions text'\n })\n})\n.then(response => response.json())\n.then(data => console.log(data))\n.catch(error => console.error(error));"
}
]
}
}
},
"components": {

157
mem0/client/main.py
View File

@@ -399,6 +399,58 @@ class MemoryClient:
capture_client_event("client.get_memory_export", self, {"keys": list(kwargs.keys())})
return response.json()
@api_error_handler
def get_custom_instructions_and_categories(self, fields: List[str]) -> Dict[str, Any]:
"""Get instructions or categories for the current project.
Args:
fields: List of field names to retrieve (e.g. ['custom_instructions', 'custom_categories'])
Returns:
Dictionary containing the requested instructions or categories.
Raises:
APIError: If the API request fails.
ValueError: If org_id or project_id are not set.
"""
if not (self.org_id and self.project_id):
raise ValueError("org_id and project_id must be set to access instructions or categories")
params = self._prepare_params({"fields": fields})
response = self.client.get(
f"/api/v1/orgs/organizations/{self.org_id}/projects/{self.project_id}/custom-instructions-and-categories/",
params=params,
)
response.raise_for_status()
capture_client_event("client.get_custom_instructions_and_categories", self, {"fields": fields})
return response.json()
@api_error_handler
def update_custom_instructions_and_categories(self, fields: Dict[str, Any]) -> Dict[str, Any]:
"""Update instructions or categories for the current project.
Args:
fields: Dictionary of fields to update
(e.g. {"custom_categories": "new instructions", "custom_categories": ["cat1"]})
Returns:
Dictionary containing the API response.
Raises:
APIError: If the API request fails.
ValueError: If org_id or project_id are not set.
"""
if not (self.org_id and self.project_id):
raise ValueError("org_id and project_id must be set to update instructions or categories")
response = self.client.post(
f"/api/v1/orgs/organizations/{self.org_id}/projects/{self.project_id}/custom-instructions-and-categories/",
json=fields,
)
response.raise_for_status()
capture_client_event("client.update_custom_instructions_and_categories", self, {"fields": list(fields.keys())})
return response.json()
def chat(self):
"""Start a chat with the Mem0 AI. (Not implemented)
@@ -604,5 +656,110 @@ class AsyncMemoryClient:
capture_client_event("async_client.reset", self.sync_client)
return {"message": "Client reset successful. All users and memories deleted."}
@api_error_handler
async def batch_update(self, memories: List[Dict[str, Any]]) -> Dict[str, Any]:
"""Batch update memories.
Args:
memories: List of memory dictionaries to update. Each dictionary must contain:
- memory_id (str): ID of the memory to update
- text (str): New text content for the memory
Returns:
str: Message indicating the success of the batch update.
Raises:
APIError: If the API request fails.
"""
response = await self.async_client.put("/v1/batch/", json={"memories": memories})
response.raise_for_status()
capture_client_event("async_client.batch_update", self.sync_client)
return response.json()
@api_error_handler
async def batch_delete(self, memories: List[Dict[str, Any]]) -> Dict[str, Any]:
"""Batch delete memories.
Args:
memories: List of memory dictionaries to delete. Each dictionary must contain:
- memory_id (str): ID of the memory to delete
Returns:
str: Message indicating the success of the batch deletion.
Raises:
APIError: If the API request fails.
"""
response = await self.async_client.request("DELETE", "/v1/batch/", json={"memories": memories})
response.raise_for_status()
capture_client_event("async_client.batch_delete", self.sync_client)
return response.json()
@api_error_handler
async def create_memory_export(self, schema: str, **kwargs) -> Dict[str, Any]:
"""Create a memory export with the provided schema.
Args:
schema: JSON schema defining the export structure
**kwargs: Optional filters like user_id, run_id, etc.
Returns:
Dict containing export request ID and status message
"""
response = await self.async_client.post("/v1/exports/", json={"schema": schema, **self._prepare_params(kwargs)})
response.raise_for_status()
capture_client_event(
"async_client.create_memory_export", self.sync_client, {"schema": schema, "keys": list(kwargs.keys())}
)
return response.json()
@api_error_handler
async def get_memory_export(self, **kwargs) -> Dict[str, Any]:
"""Get a memory export.
Args:
**kwargs: Filters like user_id to get specific export
Returns:
Dict containing the exported data
"""
response = await self.async_client.get("/v1/exports/", params=self._prepare_params(kwargs))
response.raise_for_status()
capture_client_event("async_client.get_memory_export", self.sync_client, {"keys": list(kwargs.keys())})
return response.json()
@api_error_handler
async def get_custom_instructions_and_categories(self, fields: List[str]) -> Dict[str, Any]:
if not (self.sync_client.org_id and self.sync_client.project_id):
raise ValueError("org_id and project_id must be set to access instructions or categories")
params = self.sync_client._prepare_params({"fields": fields})
response = await self.async_client.get(
f"/api/v1/orgs/organizations/{self.sync_client.org_id}/projects/{self.sync_client.project_id}/custom-instructions-and-categories/",
params=params,
)
response.raise_for_status()
capture_client_event(
"async_client.get_custom_instructions_and_categories", self.sync_client, {"fields": fields}
)
return response.json()
@api_error_handler
async def update_custom_instructions_and_categories(self, fields: Dict[str, Any]) -> Dict[str, Any]:
if not (self.sync_client.org_id and self.sync_client.project_id):
raise ValueError("org_id and project_id must be set to update instructions or categories")
response = await self.async_client.post(
f"/api/v1/orgs/organizations/{self.sync_client.org_id}/projects/{self.sync_client.project_id}/custom-instructions-and-categories/",
json=fields,
)
response.raise_for_status()
capture_client_event(
"async_client.update_custom_instructions_and_categories", self.sync_client, {"fields": list(fields.keys())}
)
return response.json()
async def chat(self):
raise NotImplementedError("Chat is not implemented yet")