How adityakumar0009/soverign-ai works
This repository appears to implement a document-ingestion API. The intended user journey is to submit a file, select a parser, process and normalize its content, validate it, and represent the result with document and related schemas. The tree also signals support for PDF, DOCX, Excel, images, PowerPoint, and text; most format-specific modules were not sampled, so their runtime connections are uncertain. The README is empty, and the shown excerpts do not establish the complete upload-to-processing call chain; uncertain links are marked dashed without evidence.
Main entry points: registry.py Parser Registry · orchestrator.py Ingestion Orchestrator · pipeline.py Processing Pipeline · parser.py PDF Parser
Components 21
API
- Application Setup
app/main.py - Upload Route
app/api/routes/upload.py - Health Route
app/api/routes/health.py
Ingestion
- Ingestion Orchestrator
app/ingestion/orchestrator.py - Parser Registry
app/ingestion/registry.py - Processing Pipeline
app/ingestion/pipeline.py - Content Normalizer
app/ingestion/normalizer.py - Document Validator
app/ingestion/validator.py
Format Parsers
- Parser Base
app/parsers/base/parser.py - PDF Parser
app/parsers/pdf/parser.py - DOCX Parser
app/parsers/docx - Excel Parser
app/parsers/excel - Image Parser
app/parsers/image - PowerPoint Parser
app/parsers/pptx - Text Parser
app/parsers/txt
Data Models
- Document Schema
app/schemas/document.py - Entity Schema
app/schemas/entity.py - Metadata Schema
app/schemas/metadata.py - Relationship Schema
app/schemas/relationship.py - Table Schema
app/schemas/table.py
Other components
- API Client
Connections 18
- Application Setup → Upload Route: registers route
app/main.py - Application Setup → Health Route: registers route
app/main.py - API Client → Upload Route: submits file
- Upload Route → Ingestion Orchestrator: hands off
- Ingestion Orchestrator → Parser Registry: uses registry
app/ingestion/orchestrator.py - Ingestion Orchestrator → Processing Pipeline: runs processing
- Processing Pipeline → Content Normalizer: normalizes
- Processing Pipeline → Document Validator: validates
- Parser Registry → PDF Parser: selects parser
- Parser Registry → DOCX Parser: selects parser
- Parser Registry → Excel Parser: selects parser
- Parser Registry → Image Parser: selects parser
- Parser Registry → PowerPoint Parser: selects parser
- Parser Registry → Text Parser: selects parser
- PDF Parser → Parser Base: implements parser
app/parsers/pdf/parser.py - PDF Parser → Document Schema: produces schema
app/parsers/pdf/parser.py - Content Normalizer → Document Schema: normalizes to
- Document Validator → Document Schema: validates schema