feat(seed): LLM dialog generator for scalable testing (#135) [отложено до лета 2026]#178
Draft
feat(seed): LLM dialog generator for scalable testing (#135) [отложено до лета 2026]#178
Conversation
Add script bin/generate_demo_dialogs that generates realistic dialogs by simulating conversation between two LLM models - one acting as a client (with behavior profile) and one as the auto service assistant. Features: - 4 client profiles: active_client, one_time_client, just_asking, urgent_client - Weighted distribution matching real customer segments - --dry-run mode for cost/time estimation - --profile flag to generate specific client type - YAML output with validation - Configurable LLM model (default: gpt-4o-mini) Closes #135 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Expand client profile library from 4 to 20 profiles: By character: - suspicious_client (40% booking, недоверчивый) - budget_client (30% booking, экономный) - vip_client (95% booking, VIP) - perfectionist_client (65% booking, перфекционист) By expertise: - tech_savvy_client (70% booking, технически грамотный) - newbie_driver (75% booking, новичок-водитель) By age/style: - elderly_client (60% booking, пожилой) - young_driver (55% booking, молодой) - busy_businessman (80% booking, занятой бизнесмен) - emotional_female_client (70% booking, эмоциональная клиентка) By situation: - referred_client (85% booking, по рекомендации) - disappointed_client (55% booking, разочарован другим сервисом) - insurance_client (90% booking, страховой случай) - pre_sale_client (70% booking, готовит к продаже) - post_purchase_client (95% booking, купил б/у авто) - corporate_client (80% booking, корпоративный) Updated PROFILES hash in script with weighted distribution. Updated README with all 20 profiles organized by category. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Generated 5 dialogs per profile using DeepSeek LLM: - Total: 100 dialogs, 1746 messages, ~1.2 MB - Profiles covered: all 20 client profiles Message distribution by profile: - active_client: 148 msgs (longest dialogs) - elderly_client: 118 msgs - suspicious_client: 110 msgs - young_driver: 48 msgs (shortest dialogs) Also updated script to use ApplicationConfig.llm_model as default instead of hardcoded gpt-4o-mini. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
dapi
changed the title
Owner
Author
|
⏸️ Отложено до лета 2026 Функционал LLM-генератора диалогов не является приоритетным на текущем этапе. Возвращаемся к нему после завершения основных фич Manager Takeover. |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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
Closes #135
Test plan
🤖 Generated with Claude Code