Skip to content

Commit a99fed7

Browse files
authored
docs updates (#216)
1 parent 64c9a31 commit a99fed7

File tree

8 files changed

+65
-5
lines changed

8 files changed

+65
-5
lines changed

docs/Roles and Access Control/1.overview.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -41,7 +41,9 @@ In the previous example, we gave access to `chat-gpt-35-turbo model` for users w
4141
"limits": {
4242
"chat-gpt-35-turbo": { //model name
4343
"minute": "100000", //number of tokens per minute
44-
"day": "10000000" //number of tokens per day
44+
"day": "10000000", //number of tokens per day
45+
"week": "10000000", //number of tokens per week
46+
"month": "10000000" //number of tokens per month
4547
}
4648
}
4749
}

docs/Roles and Access Control/2.chat-users.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -31,7 +31,9 @@ In the system configuration, you can then add rules and restrictions to manage a
3131
"limits": {
3232
"chat-gpt-35-turbo": {
3333
"minute": "200000",
34-
"day": "10000000"
34+
"day": "10000000",
35+
"week": "10000000",
36+
"month": "10000000",
3537
}
3638
}
3739
}

docs/Roles and Access Control/3.API Keys.md

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -27,7 +27,9 @@ To create and configure access control for API keys:
2727
"limits": {
2828
"chat-gpt-35-turbo": {
2929
"minute": "100000", //number of tokens per minute
30-
"day": "10000000" //number of tokens per day
30+
"day": "10000000", //number of tokens per day
31+
"week": "10000000", //number of tokens per week
32+
"month": "10000000", //number of tokens per month
3133
}
3234
}
3335
}

docs/privacy.md

Lines changed: 42 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,42 @@
1+
# PII Compliance & Privacy
2+
3+
## Introduction
4+
5+
There are two scenarios in which personal information can be exposed when working with DIAL.
6+
7+
* The first scenario takes place when an application uploads sensitive data, such as conversations or files, to DIAL via API. In this case, DIAL can capture and save conversation data in application audit logs.
8+
* The second scenario occurs when users interact with DIAL Chat, specifically when engaging with language models or applications. In this case, DIAL can capture and save conversation data in both audit logs and the user’s BLOB storage.
9+
10+
To ensure compliance with necessary Personally Identifiable Information (PII) regulations, our system offers flexibility that allows you to customize your data management strategy and ensure that your data handling practices adhere to required standards.
11+
12+
DIAL allows you to choose which logs to store (if at all), determine which data to retain, and assists in implementing necessary policies in your file storage to effectively manage sensitive resources.
13+
14+
## Applications Audit Logs
15+
16+
When a user interacts with DIAL applications programmatically using API keys, DIAL captures and records all conversation data in a designated audit log.
17+
18+
To prevent DIAL from storing such logs for a particular API key, you can enable a special flag in the API key specification in the [DIAL Core dynamic settings](https://github.com/epam/ai-dial-core/?tab=readme-ov-file#dynamic-settings):
19+
20+
```json
21+
//Example of DIAL Core dynamic settings configuration
22+
"keys":
23+
{
24+
"yourApiKey": {
25+
"secured": true
26+
}
27+
}
28+
```
29+
30+
**Important!**: It's important to be aware that custom applications can upload resources, like conversations or files, to DIAL. If this is the case, the logic within the custom application that carries out such actions is responsible for handling these resources. To manage them, you can additionally implement Time To Live (TTL) or other policies in your file storage. For more details on this topic, please refer to the [File Storage Policies](#file-storage-policies) section.
31+
32+
## BLOB Storage
33+
34+
When a user interacts with DIAL Chat using a JSON Web Token (JWT), DIAL captures and records all conversation data in a designated audit log and **also** stores it as JSON files in the user's BLOB storage.
35+
36+
**Important!**: It's important to be aware that even if a user deletes all conversation data from the BLOB storage the data will still be retained in the audit logs. Because DIAL Chat uses JWT for user authentication, rather than API keys, the information will inevitably be saved in BLOB storage.
37+
38+
## File Storage Policies
39+
40+
To manage your resource uploaded to the BLOB storage, you can configure policies, which operate independently from DIAL. BLOB storage policies can utilize functionalities such as cloud lambdas to establish TTLs for specific file types.
41+
42+
To facilitate the enforcement of such policies, DIAL can add metadata to files stored in the BLOB storage. This metadata assists in the application of storage policies, ensuring effective management of your resources.

docs/supported-models.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -7,7 +7,7 @@ You can use [DIAL SDK](https://github.com/epam/ai-dial-sdk) to create custom mod
77
| Vendor | Models |
88
| :-- | :-- |
99
| AI21| ai21.j2-grande-instruct, ai21.j2-jumbo-instruct |
10-
| Amazon| amazon.titan-tg1-large, amazon.titan-embed-text-v1, amazon.titan-embed-text-v2:0, amazon.titan-embed-image-v1 |
10+
| Amazon| amazon.titan-tg1-large, amazon.titan-embed-text-v1, amazon.titan-embed-text-v2:0, amazon.titan-embed-image-v1, amazon.nova-pro-v1, amazon.nova-lite-v1, amazon.nova-micro-v1|
1111
| Anthropic| anthropic.claude-instant-v1, anthropic.claude-v2, anthropic.claude-v2-1, anthropic.claude-v3-opus, anthropic.claude-v3-haiku, anthropic.claude-3-5-haiku-20241022-v1, anthropic.claude-v3-sonnet, anthropic.claude-v3-5-sonnet, anthropic.claude-3-5-sonnet-20241022-v2, anthropic.claude |
1212
| Cohere| cohere.command-text-v14 |
1313
| Databricks| databricks-bge-large-en, databricks-llama-2-70b-chat, databricks-mixtral-8x7b-instruct, databricks-dbrx-instruct |

docs/tutorials/rate-limits-users.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -33,7 +33,7 @@ For example purposes, lets configure rate limits for AI DIAL Chat users with use
3333
}
3434
}
3535
```
36-
5. In AI DIAL Core configuration file, in the `roles` section, add your role ("azure-group-name") and limits for it. Refer to [configuration example](https://github.com/epam/ai-dial-core/blob/9d7e3ba8380ffea3b9b6a7ccd65a96f024e842e3/sample/aidial.config.json#L191) for example purposes. For `roles.<role_name>.limits` you can configure `minute` (total tokens per minute limit sent to the model, managed via floating window approach for well-distributed rate limiting. If it's not set the default value is unlimited) or `day` (total tokens per day limit sent to the model, managed via floating window approach for balanced rate limiting. If it's not set the default value is unlimited).
36+
5. In AI DIAL Core configuration file, in the `roles` section, add your role ("azure-group-name") and limits for it. Refer to [configuration example](https://github.com/epam/ai-dial-core/blob/9d7e3ba8380ffea3b9b6a7ccd65a96f024e842e3/sample/aidial.config.json#L191) for example purposes. For `roles.<role_name>.limits` you can configure `minute` (total tokens per minute limit sent to the model, managed via floating window approach for well-distributed rate limiting. If it's not set the default value is unlimited), `day` (total tokens per day limit sent to the model, managed via floating window approach for balanced rate limiting. If it's not set the default value is unlimited), `week` and `month` accordingly.
3737

3838
> The `default` role applies in case other roles are not configured.
3939
> In case the same user has different roles with different limits, the role with the higher limit is an effective role.

docs/video demos/demos/dial-chathub.md

Lines changed: 11 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3,3 +3,14 @@
33
https://youtu.be/IG0HawQuU-w
44

55
DIAL ChatHub is an example of a flow orchestrator that combines several applications and models into one unified access point. ChatHub can automatically route prompts to one of several agents: text-to-text applications, text-to-image applications, vision-to-text applications, DIAL RAG, or DIAL Web RAG. It is just an example of how you can provide a unified access point to several different applications or models within your ecosystem.
6+
7+
https://youtu.be/8Npbd0rESPI
8+
9+
In this video, we explore the innovative prompt routing engine, ChatHub 2.0, that revolutionizes how users interact with AI tools. Learn how this intelligent system:
10+
11+
* Provides a single entry point for all your GenAI chat needs
12+
* Uses a GPT-4o based orchestrator to route prompts to the best tool
13+
* Seamlessly integrates existing DIAL applications like RAG
14+
* Leverages advanced models such as Gemini 1.5 Pro with Search
15+
* Parallelizes searches and chains agents for efficient responses
16+
* Combines text, data, and image generation capabilities

sidebars.js

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -7,6 +7,7 @@ const sidebars = {
77
label: 'Home', // sidebar label
88
},
99
'quick-start',
10+
'privacy',
1011
'architecture',
1112
'supported-models',
1213
{

0 commit comments

Comments
 (0)