Hugging Face's logo

Spaces:
milwright
/
cloze-reader
Sleeping

App Files Community
cloze-reader / src /aiService.js

Commit History

fix leaderboard initials entry readability with improved contrast and light background
e4e854a

milwright commited on

simplify game mechanics and consolidate to single AI model
86f177c

milwright Claude commited on

fix: prevent numbers from being selected as blank words
4f630fa

milwright commited on

fix ui issues: larger chat icons and proper book contextualization
37a7438

milwright commited on

simplify ui language and integrate dual gemma models
fed2199

milwright commited on

Add local LLM integration support
56287e6

milwright commited on

Fix difficulty progression and batch processing issues
1505067

milwright commited on

Optimize AI word selection to reduce fallback rate
de88fe1

milwright commited on

Fix level progression and word validation issues
eeb2654

milwright commited on

Improve content filtering and word validation
e26fd03

milwright commited on

Add word length validation to batch processing
d5848fc

milwright commited on

Implement level-aware word selection and performance optimizations
9e0745a

milwright commited on

fix level progression and optimize ai word selection
440ff08

milwright commited on

fix api key loading and enhance ai prompts
a918c3e

milwright commited on

fix gemma-3 api compatibility by removing system messages
559e227

milwright commited on

improve prompting tactics and retry logic for multiple tasks
b97292d

milwright commited on

fix: improve UI elements and reduce input field width
1df4d55

milwright commited on

fix: improve chat interface and error handling
3ca518f

milwright commited on

feat: implement batch API processing for both passages
cd16e7b

milwright commited on

fix: improve word selection and passage quality filtering
b37374a

milwright commited on

fix: clean up AI response processing and improve UI
b451c6e

milwright commited on

feat: implement two-passage system to avoid OpenRouter rate limits
106bc8b

milwright commited on

debug: add API response logging to identify contextualization issue after first round
cca074a

milwright commited on

fix: clean contextualization artifacts and ensure passages start properly
4f8034e

milwright commited on

fix: clean AI responses to remove JSON artifacts and markdown formatting
dfcab2f

milwright commited on

fix: resolve HF dataset loading errors and improve word selection
6839f0d

milwright commited on

fix: provide target word to AI while preventing disclosure
13ed840

milwright commited on

fix: add robust API error handling for data.choices undefined
9c565fb

milwright commited on

fix: improve word selection and remove markdown formatting
b874e04

milwright commited on

fix: implement constrained prompts matching tool-calling behavior
18f0d59

milwright commited on

refactor: reduce technical debt by removing unused methods and tool definitions
6e0c330

milwright commited on

refactor: proper separation of concerns - aiService as API wrapper, conversationManager builds prompts
2b8632d

milwright commited on

fix: make chat responses direct and concise like proper clues
7a48149

milwright commited on

fix: replace tool-calling with simple direct prompts like main branch
a261c04

milwright commited on

fix: prevent AI hallucination and remove problematic tool_choice parameter
243e743

milwright commited on

fix: correct tool-calling format and contextualization truncation
78a1f73

milwright commited on

fix: restore working chat and contextualization logic
09d70c8

milwright commited on

fix: restore tool-calling chat prompt engineering
25876df

milwright commited on

fix: add runtime API key checking and comprehensive debugging
7d032f3

milwright commited on

fix: restore missing AI methods for word selection and chat
47a380e

milwright commited on

feat: add clean aiService.js without hardcoded keys
63a5bc1

milwright commited on