# jayeshvadje/telecom-support-intelligence architecture

> Architecture diagram of the GitHub repository jayeshvadje/telecom-support-intelligence, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-06). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/jayeshvadje/telecom-support-intelligence
- Repository: https://github.com/jayeshvadje/telecom-support-intelligence

## Overview

This repository implements a telecom-policy RAG service plus a document-indexing workflow. Policy documents are parsed, chunked, enriched, embedded, and upserted to Pinecone; API queries are embedded and searched, then Gemini generates grounded answers with source metadata. A separate evaluator scores answers. The source excerpts establish these runtime calls, but no README was provided and app.py was not excerpted, so its role is not represented.

Main entry points: [`ingestion/pipeline.py`](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/pipeline.py) (Ingestion pipeline), [`generation/rag_chain.py`](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/generation/rag_chain.py) (RAG answer chain), [`retrieval/pinecone_retriever.py`](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/retrieval/pinecone_retriever.py) (Semantic retriever), [`evaluation/evaluator.py`](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/evaluation/evaluator.py) (RAG evaluator)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_ingestion["Policy ingestion"]
  node_policy_docs["Policy documents"]
  node_ingest_pipeline["Ingestion pipeline<br/>[pipeline.py]"]
  node_docx_loader["DOCX loader<br/>[docx_loader.py]"]
  node_text_cleaner["Text cleaner<br/>[text_cleaner.py]"]
  node_policy_chunker["Policy chunker<br/>[policy_chunker.py]"]
  node_metadata_extractor["Metadata extractor<br/>[extractor.py]"]
end

subgraph group_index["Vector indexing"]
  node_gemini_embedder["Document embedder<br/>[gemini_embedder.py]"]
  node_pinecone_manager["Index manager"]
  node_pinecone_store[("Pinecone index")]
end

subgraph group_serving["Query service"]
  node_query_api["Query API<br/>[main.py]"]
  node_rag_chain["RAG answer chain<br/>[rag_chain.py]"]
  node_query_retriever["Semantic retriever"]
end

subgraph group_evaluation["Answer evaluation"]
  node_evaluator["RAG evaluator<br/>[evaluator.py]"]
end

node_query_user(("Support user"))
node_gemini_service["Google Gemini"]

node_policy_docs -->|"provides files"| node_ingest_pipeline
node_ingest_pipeline -->|"parses documents"| node_docx_loader
node_docx_loader -->|"cleans text"| node_text_cleaner
node_ingest_pipeline -->|"creates chunks"| node_policy_chunker
node_ingest_pipeline -->|"enriches chunks"| node_metadata_extractor
node_ingest_pipeline -->|"embeds chunks"| node_gemini_embedder
node_gemini_embedder -->|"requests embeddings"| node_gemini_service
node_ingest_pipeline -->|"upserts chunks"| node_pinecone_manager
node_pinecone_manager -->|"writes vectors"| node_pinecone_store
node_query_user -->|"submits query"| node_query_api
node_query_api -->|"requests answer"| node_rag_chain
node_rag_chain -->|"searches policies"| node_query_retriever
node_query_retriever -->|"embeds query"| node_gemini_service
node_query_retriever -->|"queries vectors"| node_pinecone_store
node_rag_chain -->|"generates answer"| node_gemini_service
node_evaluator -->|"gets answer"| node_rag_chain
node_evaluator -->|"gets context"| node_query_retriever
node_evaluator -->|"scores response"| node_gemini_service

click node_policy_docs "https://github.com/jayeshvadje/telecom-support-intelligence/tree/main/ingestion/data"
click node_ingest_pipeline "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/pipeline.py"
click node_docx_loader "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/loaders/docx_loader.py"
click node_text_cleaner "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/processors/text_cleaner.py"
click node_policy_chunker "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/chunking/policy_chunker.py"
click node_metadata_extractor "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/metadata/extractor.py"
click node_gemini_embedder "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/embeddings/gemini_embedder.py"
click node_pinecone_manager "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/vectorstore/pinecone_manager.py"
click node_query_api "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/main.py"
click node_rag_chain "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/generation/rag_chain.py"
click node_query_retriever "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/retrieval/pinecone_retriever.py"
click node_evaluator "https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/evaluation/evaluator.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_policy_docs,node_ingest_pipeline,node_docx_loader,node_text_cleaner,node_policy_chunker,node_metadata_extractor,node_query_user toneBlue
class node_gemini_embedder,node_pinecone_manager,node_pinecone_store toneAmber
class node_query_api,node_rag_chain,node_query_retriever toneMint
class node_evaluator toneRose
class node_gemini_service toneIndigo
```

## Components

### Policy ingestion

- [Policy documents](https://github.com/jayeshvadje/telecom-support-intelligence/tree/main/ingestion/data): `ingestion/data`
- [Ingestion pipeline](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/pipeline.py): `ingestion/pipeline.py`
- [DOCX loader](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/loaders/docx_loader.py): `ingestion/loaders/docx_loader.py`
- [Text cleaner](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/processors/text_cleaner.py): `ingestion/processors/text_cleaner.py`
- [Policy chunker](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/chunking/policy_chunker.py): `ingestion/chunking/policy_chunker.py`
- [Metadata extractor](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/metadata/extractor.py): `ingestion/metadata/extractor.py`

### Vector indexing

- [Document embedder](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/embeddings/gemini_embedder.py): `ingestion/embeddings/gemini_embedder.py`
- [Index manager](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/ingestion/vectorstore/pinecone_manager.py): `ingestion/vectorstore/pinecone_manager.py`
- **Pinecone index**

### Query service

- [Query API](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/main.py): `main.py`
- [RAG answer chain](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/generation/rag_chain.py): `generation/rag_chain.py`
- [Semantic retriever](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/retrieval/pinecone_retriever.py): `retrieval/pinecone_retriever.py`

### Answer evaluation

- [RAG evaluator](https://github.com/jayeshvadje/telecom-support-intelligence/blob/main/evaluation/evaluator.py): `evaluation/evaluator.py`

### Other components

- **Support user**
- **Google Gemini**

## Connections

- Policy documents → Ingestion pipeline: provides files (evidence: `ingestion/pipeline.py`)
- Ingestion pipeline → DOCX loader: parses documents (evidence: `ingestion/pipeline.py`)
- DOCX loader → Text cleaner: cleans text (evidence: `ingestion/loaders/docx_loader.py`)
- Ingestion pipeline → Policy chunker: creates chunks (evidence: `ingestion/pipeline.py`)
- Ingestion pipeline → Metadata extractor: enriches chunks (evidence: `ingestion/pipeline.py`)
- Ingestion pipeline → Document embedder: embeds chunks (evidence: `ingestion/pipeline.py`)
- Document embedder → Google Gemini: requests embeddings (evidence: `ingestion/embeddings/gemini_embedder.py`)
- Ingestion pipeline → Index manager: upserts chunks (evidence: `ingestion/pipeline.py`)
- Index manager → Pinecone index: writes vectors (evidence: `ingestion/vectorstore/pinecone_manager.py`)
- Support user → Query API: submits query (evidence: `main.py`)
- Query API → RAG answer chain: requests answer (evidence: `main.py`)
- RAG answer chain → Semantic retriever: searches policies (evidence: `generation/rag_chain.py`)
- Semantic retriever → Google Gemini: embeds query (evidence: `retrieval/pinecone_retriever.py`)
- Semantic retriever → Pinecone index: queries vectors (evidence: `retrieval/pinecone_retriever.py`)
- RAG answer chain → Google Gemini: generates answer (evidence: `generation/rag_chain.py`)
- RAG evaluator → RAG answer chain: gets answer (evidence: `evaluation/evaluator.py`)
- RAG evaluator → Semantic retriever: gets context (evidence: `evaluation/evaluator.py`)
- RAG evaluator → Google Gemini: scores response (evidence: `evaluation/evaluator.py`)
