Dev ai #264
Open
Dev ai #264
+711
−337
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
… `TextDeepSeekModel`
Reviewer's Guide by SourceryThis pull request introduces the ability to configure the AI text model used by the application. It also adds support for the DeepSeek model. Class diagram for AI text service changesclassDiagram
class IAITextService {
<<interface>>
+init(IAITextModelOpts)
}
class MainAITextService {
-_modelType: TextModelType
+init(IAITextModelOpts)
-createModel()
}
class TextModelType {
<<enumeration>>
GPT
DeepSeek
}
MainAITextService ..|> IAITextService
note for MainAITextService "Added DeepSeek model support"
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey @Bistard - I've reviewed your changes - here's some feedback:
Overall Comments:
- The MainAITextService returns GPTModel for both TextModelType.GPT and TextModelType.DeepSeek. If DeepSeek should have different behavior, it needs its own model implementation.
Here's what I looked at during the review
- 🟡 General issues: 2 issues found
- 🟢 Security: all looks good
- 🟢 Testing: all looks good
- 🟡 Complexity: 1 issue found
- 🟢 Documentation: all looks good
Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
| @@ -53,6 +50,9 @@ export class MainAITextService extends Disposable implements IAITextService { | |||
| switch (this._modelType) { | |||
| case TextModelType.GPT: | |||
| return new GPTModel(); | |||
There was a problem hiding this comment.
issue (bug_risk): DeepSeek model type falls back to GPTModel without warning
This could cause unexpected behavior as users expecting DeepSeek will silently get GPT instead. Consider implementing DeepSeek model or throwing an error if not yet supported.
| @@ -145,7 +145,7 @@ interface IJsonSchemaForString extends IJsonSchemaBase<'string'> { | |||
| /** The maximum length of the string. */ | |||
| maxLength?: number; | |||
|
|
|||
| /** The predefined format of the string. Example: 'email', 'phone number', 'post address' etc. */ | |||
| /** The predefined format of the string (written Regular Expression). Example: 'email', 'phone number', 'post address' etc. */ | |||
There was a problem hiding this comment.
issue (typo): Incorrect documentation for JSON Schema format field
The format field in JSON Schema is for predefined format identifiers (like 'date-time', 'email', 'uri'), not for regular expressions. The pattern field is used for regex patterns.
Suggested change
| /** The predefined format of the string (written Regular Expression). Example: 'email', 'phone number', 'post address' etc. */ | |
| /** The predefined format identifier as defined by JSON Schema. Example: 'date-time', 'email', 'uri', 'hostname' etc. */ |
| @@ -1,4 +1,5 @@ | |||
| import { CollapseState } from "src/base/browser/basic/dom"; | |||
There was a problem hiding this comment.
issue (complexity): Consider simplifying the configuration registration by creating a helper function to reduce boilerplate and improve maintainability.
The configuration registration can be simplified by introducing a helper function while maintaining separation of concerns. For example:
// Helper function
function registerConfigSection<T extends string>(
sectionName: string,
properties: Record<string, {
type: string,
default: T,
enum?: T[]
}>
) {
return createRegister(
RegistrantType.Configuration,
sectionName,
(registrant) => {
registrant.registerConfigurations({
id: sectionName,
properties: {
[sectionName]: {
type: 'object',
properties
}
}
});
}
);
}
// Usage for AI config
export const sharedApplicationConfigurationRegister = registerConfigSection(
'ai',
{
textModel: {
type: 'string',
default: TextModelType.DeepSeek,
enum: [TextModelType.DeepSeek, TextModelType.GPT]
}
}
);This approach:
- Reduces boilerplate while keeping configurations separate
- Makes adding new configuration sections more concise
- Maintains type safety
- Preserves the existing functionality
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary by Sourcery
Add support for configurable AI text models.
New Features:
Tests: