Docs: Add DepthAPI project overview and extraction libs

README.md

@@ -16,7 +16,16 @@ size_categories:
 
 The **DepthAPI Technical Corpus** is a curated, high-quality retrieval corpus designed for modern RAG (Retrieval-Augmented Generation) systems. It features clean, aggressively normalized technical documentation, code snippets, engineering post-mortems, and system design literature.
 
-This dataset was explicitly built to serve as the local ground-truth for the [DepthAPI](https://github.com/sanjeevafk/depthapi) project—an enterprise-grade, declarative RAG pipeline built with asynchronous concurrency and Supabase Vector embeddings.
+This dataset was explicitly built to serve as the local ground-truth for the [DepthAPI](https://github.com/sanjeevafk/depthapi) project. 
+
+### About the DepthAPI Project
+DepthAPI is an enterprise-grade, declarative RAG pipeline built with asynchronous concurrency and Supabase Vector embeddings. It aims to systematize RAG data ingestion by moving away from hardcoded scripts to a configuration-driven architecture, ensuring maximum throughput, observability, and resilience. 
+
+Key features of DepthAPI include:
+- **Plugin-based Architecture**: Easily extendable ingestion strategies via declarative configurations.
+- **Async Concurrency**: High-throughput processing using bounded, multi-document orchestration.
+- **Namespace Isolation**: Precise control over data organization within the vector store.
+- **Idempotency**: Safe, hash-based upserts to avoid duplication and state corruption.
 
 ## Included Data Sources
 
@@ -29,7 +38,11 @@ The corpus aggregates multiple highly-valued technical domains, explicitly isola
 
 ## Ingestion Pipeline Details
 
-The corpus is powered by the **DepthAPI Declarative Ingestion Pipeline**, a robust architectural layer that moves away from hardcoded scripts to a configuration-driven (`config.yaml`) approach. 
+The corpus is powered by the **DepthAPI Declarative Ingestion Pipeline**, moving from manual scripts to a `config.yaml` driven approach. 
+
+During the ingestion of this corpus, we heavily utilized the following open-source extraction libraries:
+- [**Scrapling**](https://github.com/D4Vinci/Scrapling): Used for live technical documentation crawling and structured HTML extraction from documentation websites.
+- [**opendataloader-pdf**](https://github.com/opendataloader-project/opendataloader-pdf): Used for PDF extraction when ingesting book-like and document-style technical sources into normalized markdown/text blocks.
 
 Key features of the pipeline used to create this dataset:
 - **Declarative Sources:** Ingestion configs declare `LocalDirSource` with strict glob matching.