# Changelog of TikTok Shop Reviews & Product Sentiment Analyzer (`nexascout/tiktok-shop-reviews-sentiment-analyzer`) Actor

- **URL**: https://apify.com/nexascout/tiktok-shop-reviews-sentiment-analyzer/changelog.md
- **Full Actor documentation**: https://apify.com/nexascout/tiktok-shop-reviews-sentiment-analyzer.md

## Changelog

### 0.1.5 — Store discovery README

- Rebuilt the Store README around review-scraping and customer-feedback outcomes.
- Replaced the obsolete `$114` example with the verified post-fix result where unavailable price remains null.
- Added transparent billing, long-tail FAQ, and complete NexaScout Suite links.

### 0.1.4 — Product-price provenance fix

- Never interpret currency amounts inside review text or generic page copy as the current product price.
- Penalize review-shaped objects during product-object selection and ignore their generic `price` fields.
- Keep price null when TikTok exposes reviews but no structured product price or product metadata.
- Added regression coverage for the live `$114` order complaint versus the verified `$45.99` GOOLOO listing price.

### 0.1.3 — Topic summary accuracy fix

- Prevent neutral-only topics from appearing under `topPositiveTopics`.
- Require actual positive or negative review counts before promoting a topic into the respective top-topic list.
- Added `order_fulfillment` for quantity, receipt, and order complaints.
- Removed ambiguous standalone `charge` matching from `battery_charging` to avoid confusing billing complaints with product charging performance.

### 0.1.2 — Review quality and sentiment fix

- Reject generic TikTok promotion labels that only contain an ID and text.
- Require rating, reviewer, timestamp, verification, or explicit review fields before accepting sparse network objects as reviews.
- Detect explicit fulfillment complaints even when a buyer submitted a high star rating.
- Added regression coverage for live `Back to School`, `Free shipping`, `Flash sale`, quantity-shortfall, and disappointing-casing examples.

### 0.1.1 — Input schema validation fix

- Added the required description for the nested `productUrls[].url` input field.
- Added the missing description for `navigationTimeoutSecs`.
- Fixed Apify input-schema validation and cloud build deployment.

### 0.1.0 — Store launch candidate

- Added direct product URL and keyword-based TikTok Shop product selection.
- Added owned public review extraction from TikTok page state, network responses, scripts, and visible review elements.
- Added normalized review fields, stable deduplication, and per-product review caps.
- Added deterministic English sentiment scoring with negation handling.
- Added product-level praise, complaint, rating-distribution, and recurring-topic summaries.
- Added MCP-friendly input, output, and dataset schemas.
- Added limited permissions, a lightweight default input, source diagnostics, and zero external Actor/API dependencies.
- Added 25 automated extraction, normalization, scoring, and sentiment tests.