Pipeline Architecture
======================

The TEXTRA-IA Research pipeline is composed of several interconnected components that work together to process, analyze, and synthesize scientific publications. This document details each component and their interactions.

.. figure:: _static\images\pipeline_updated.png
   :width: 100%
   :align: center
   :alt: TEXTRA-IA Research Pipeline Architecture

   TEXTRA-IA Research Pipeline Architecture Diagram


Input Processing
-----------------

The system accepts three types of input:

1. **Audio Input**

   - Processed through audio-to-text conversion
   - Converted text feeds into the main context processing pipeline

2. **Text Input**

   - Direct text input from scientific publications
   - Feeds directly into context processing

3. **Image Input**

   - Undergoes specialized image processing
   - Generates image descriptions using computer vision
   - Descriptions are integrated with textual context

.. figure:: _static/images/inputs.png
   :width: 100%
   :align: center
   :alt: Input Processing

   Input Processing


Context and Synthesis
----------------------

The core processing pipeline consists of:

1. **Context Processing**

   - Aggregates inputs from multiple sources
   - Integrates text, transcribed audio, and image descriptions
   - Produces PDF output for archival
   - Feeds into the multi-agent system

2. **Multi-Agent System**

   - Coordinates analysis across specialized agents
   - Manages task distribution and aggregation
   - Ensures coherent processing flow

3. **Synthesis**

   - Generates final outputs based on multi-agent processing
   - Creates comprehensive research summaries and analyses


Knowledge Base Creation
------------------------

The knowledge base creation process involves:

1. **Batch Processing**

   - Handles multiple PDF documents simultaneously
   - Extracts structured information
   - Stores in database format

2. **Database Structure**

   - PDF metadata storage
      * Clé (hashlib)
      * Title
      * References
      * Notes/clés
      * Date
      * Path
      * Vectorized (boolean)
   - Figure handling
      * ID
      * Image description
      * Type

3. **Vectorization**

   - Converts processed text into vector representations
   - Utilizes LVM finetuned models
   - Enables efficient similarity search and analysis

.. figure:: _static/images/database.png
   :width: 100%
   :align: center
   :alt: Knowledge Base Creation

   Knowledge Base Creation


Concepts Graph Processing
--------------------------

The concept graph generation consists of two main phases:

1. **Processing Phase**

   - Vector DB integration
   - NER (Named Entity Recognition)
   - Relation extraction
   - Concept clustering

2. **Graph Creation Phase**

   - Node creation from extracted concepts
   - Edge definition between related concepts
   - Edge weighting based on relationship strength
   - Final graph generation

.. figure:: _static/images/concept.png
   :width: 100%
   :align: center
   :alt: Concepts Graph Processing

   Concepts Graph Processing


Research Synthesis System
--------------------------

The research synthesis component operates in two stages:

1. **Retrieval Stage**

   - Vector search in the knowledge base
   - Context ranking for relevance
   - Extraction of relevant excerpts
   - Topic analysis for categorization

2. **Synthesis Stage**

   - Information consolidation
   - Formatting for output
   - Final synthesis generation

.. figure:: _static/images/research.png
   :width: 100%
   :align: center
   :alt: Research Synthesis System

   Research Synthesis System


Timeline and Trend Analysis
----------------------------

A dedicated system for temporal analysis includes:

1. **Data Sources**

   - Vector DB for concept information
   - Temporal data for chronological analysis
   - Publication metadata for contextual information

2. **Analysis Components**

   - Temporal clustering of research topics
   - Keyword evolution tracking
   - Citation pattern analysis

3. **Output Generation**

   - Growth trajectory visualization
   - Trend scoring and analysis
   - Timeline generation
   - Trend visualization
   - Emerging concepts identification

.. figure:: _static/images/trends.png
   :width: 100%
   :align: center
   :alt: Timeline and Trend Analysis

   Timeline and Trend Analysis


Integration Points
-------------------

The system maintains several critical integration points:

1. **Data Flow**

   - Seamless transfer between components
   - Consistent data format maintenance
   - Error handling and recovery

2. **Vector Database**

   - Central repository for processed information
   - Enables efficient retrieval and analysis
   - Maintains relationships between concepts

3. **Output Generation**

   - Multiple output formats (PDF, visualizations, summaries)
   - Customizable based on user needs
   - Integration with existing research workflows