From e9a5a55eb00b191cfd2fb3795a06f982e65cf0e4 Mon Sep 17 00:00:00 2001
From: Tan Le <tle@gitlab.com>
Date: Fri, 5 Jan 2024 23:26:25 +1100
Subject: [PATCH] Update AI Gateway environment variables in doc
---
doc/development/ai_features/index.md | 17 ++++++++---------
doc/development/code_suggestions/index.md | 20 ++++++++++----------
2 files changed, 18 insertions(+), 19 deletions(-)
diff --git a/doc/development/ai_features/index.md b/doc/development/ai_features/index.md
index 4c69f26c2dee0..87c82f85a5af9 100644
--- a/doc/development/ai_features/index.md
+++ b/doc/development/ai_features/index.md
@@ -181,10 +181,9 @@ Therefore, a different setup is required from the [SaaS-only AI features](#test-
1. Ensure that the following environment variables are set in the `.env` file:
```shell
- AUTH_BYPASS_EXTERNAL=true
- ANTHROPIC_API_KEY="[REDACTED]" # IMPORTANT: Ensure you use Corp account. See https://gitlab.com/gitlab-org/gitlab/-/issues/435911#note_1701762954.
- PALM_TEXT_MODEL_NAME=text-bison
- PALM_TEXT_PROJECT="[REDACTED]"
+ AIGW_AUTH__BYPASS_EXTERNAL=true
+ ANTHROPIC_API_KEY="[REDACTED]" # IMPORTANT: Ensure you use Corp account. See https://gitlab.com/gitlab-org/gitlab/-/issues/435911#note_1701762954.
+ AIGW_VERTEX_TEXT_MODEL__PROJECT="[REDACTED]"
```
1. Run `poetry run ai_gateway`.
@@ -215,7 +214,7 @@ Therefore, a different setup is required from the [SaaS-only AI features](#test-
1. Create a dummy access token via `gdk rails console` OR skip this step and setup GitLab or Customer Dot as OIDC provider (See the following section):
```ruby
- # Creating dummy token, and this will work as long as `AUTH_BYPASS_EXTERNAL=true` in AI Gateway.
+ # Creating dummy token, and this will work as long as `AIGW_AUTH__BYPASS_EXTERNAL=true` in AI Gateway.
::Ai::ServiceAccessToken.create!(token: 'dummy', expires_at: 1.month.from_now)
```
@@ -272,9 +271,9 @@ Therefore, a different setup is required from the [SaaS-only AI features](#test-
1. Additionally, ensure that the following environment variables are set in the `.env` file:
```shell
- GITLAB_URL="http://gdk.test:3000/"
- GITLAB_API_URL="http://gdk.test:3000/api/v4/"
- AUTH_BYPASS_EXTERNAL=False
+ AIGW_GITLAB_URL="http://gdk.test:3000/"
+ AIGW_GITLAB_API_URL="http://gdk.test:3000/api/v4/"
+ AIGW_AUTH__BYPASS_EXTERNAL=False
```
1. Restart AI Gateway.
@@ -290,7 +289,7 @@ Therefore, a different setup is required from the [SaaS-only AI features](#test-
### Use Customer Dot as OIDC provider in AI Gateway
1. AI Gateway:
- 1. Ensure `CUSTOMER_PORTAL_BASE_URL` in the `.env` file points to your Customer Dot URL.
+ 1. Ensure `AIGW_CUSTOMER_PORTAL_BASE_URL` in the `.env` file points to your Customer Dot URL.
1. Restart
## Experimental REST API
diff --git a/doc/development/code_suggestions/index.md b/doc/development/code_suggestions/index.md
index f282dbb122399..2bf3666443753 100644
--- a/doc/development/code_suggestions/index.md
+++ b/doc/development/code_suggestions/index.md
@@ -12,7 +12,7 @@ The recommended setup for locally developing and debugging Code Suggestions is t
- IDE Extension (e.g. VS Code Extension)
- Main application configured correctly
-- [Model gateway](https://gitlab.com/gitlab-org/modelops/applied-ml/code-suggestions/ai-assist)
+- [AI Gateway](https://gitlab.com/gitlab-org/modelops/applied-ml/code-suggestions/ai-assist)
This should enable everyone to see locally any change in an IDE being sent to the main application transformed to a prompt which is then sent to the respective model.
@@ -29,21 +29,21 @@ This should enable everyone to see locally any change in an IDE being sent to th
1. Run `bundle exec rails c` to start a Rails console
1. Call `Feature.enable(:code_suggestions_tokens_api)` from the console
1. Run the GDK with ```export CODE_SUGGESTIONS_BASE_URL=http://localhost:5052```
-1. [Setup Model Gateway](https://gitlab.com/gitlab-org/modelops/applied-ml/code-suggestions/ai-assist#how-to-run-the-server-locally)
+1. [Setup AI Gateway](https://gitlab.com/gitlab-org/modelops/applied-ml/code-suggestions/ai-assist#how-to-run-the-server-locally)
1. Build tree sitter libraries ```poetry run scripts/build-tree-sitter-lib.py```
- 1. Extra .env Changes for all debugging insights
- 1. LOG_LEVEL=DEBUG
- 1. LOG_FORMAT_JSON=false
- 1. LOG_TO_FILE=true
+ 1. Extra .env changes for all debugging insights
+ 1. `AIGW_LOGGING__LEVEL=DEBUG`
+ 1. `AIGW_LOGGING__FORMAT_JSON=false`
+ 1. `AIGW_LOGGING__TO_FILE=true`
1. Watch the new log file ```modelgateway_debug.log``` , e.g. ```tail -f modelgateway_debug.log | fblog -a prefix -a suffix -a current_file_name -a suggestion -a language -a input -a parameters -a score -a exception```
-### Setup instructions to use staging Model Gateway
+### Setup instructions to use staging AI Gateway
-When testing interactions with the Model Gateway, you might want to integrate your local GDK
-with the deployed staging Model Gateway. To do this:
+When testing interactions with the AI Gateway, you might want to integrate your local GDK
+with the deployed staging AI Gateway. To do this:
1. You need a [cloud staging license](../../user/project/repository/code_suggestions/self_managed.md#upgrade-gitlab) that has the Code Suggestions add-on, because add-ons are enabled on staging. Drop a note in the `#s_fulfillment` internal Slack channel to request an add-on to your license. See this [handbook page](https://about.gitlab.com/handbook/developer-onboarding/#working-on-gitlab-ee-developer-licenses) for how to request a license for local development.
-1. Set environment variables to point customers-dot to staging, and the Model Gateway to staging:
+1. Set environment variables to point customers-dot to staging, and the AI Gateway to staging:
```shell
export GITLAB_LICENSE_MODE=test
--
GitLab