[DMP 2024]:

[piyushmittaltech]

Ticket Contents

Protean Overview

Protean eGov Technologies Limited provides IT services. The Company offers citizen services, e-governance solutions, system integration, business process re-engineering, data center colocation, and IT consulting services for citizens, corporations and the Government. We have been modernizing the direct tax infrastructure, providing a tax identity to citizens and corporates (issuance of PAN card), strengthening the old-age social security infrastructure (National Pension System NPS & Atal Pension Yojna - APY), promoting financial inclusion by contributing to the India Stack by enrolling citizens for National Identity and enabling the BFSI sector by providing Aadhaar-based identity authentication and e-Sign services

Benefits of this Project

• Improved agricultural productivity and sustainability through AI-driven recommendations.
• Enhanced accessibility with multilingual support for farmers across India.
• Streamlined development and integration process with well-documented and tested APIs.

Description

Use publicly available data to build a predictable model to provide personalised prescriptions, dietary, nutritionally, etc predictions.

The project involves developing robust APIs for an Agronomy app that utilizes AI and ML on historical data to offer personalized agricultural advice and predictive analytics. The APIs will enable features such as disease prediction, seed and fertilizer recommendations, nutrition advice, and increased crop yield insights, with multilingual support for all Indian regional languages.

Personalized Agri-Advice API: As an agronomist, I want to access personalized agricultural advice through API endpoints. Tasks:

• [x] Develop API endpoints to provide tailored recommendations based on crop type, soil conditions, and climate.
• [x] Integrate AI/ML models to generate personalized agri-advice.
• [x] Test API endpoints with diverse scenarios.

Disease Prediction and Cure Advisory API: As a farmer, I want an API to predict crop diseases and recommend cures. Tasks:

• [x] Implement API endpoints for disease prediction using AI/ML algorithms.
• [x] Provide advisory recommendations for disease management.
• [x] Test API accuracy with historical disease data.

Seed and Fertilizer Recommendations API: As a user, I want an API to suggest suitable seeds and fertilizers based on my farming requirements. Tasks:

• [x] Develop API endpoints for seed and fertilizer recommendations using predictive analytics.
• [x] Integrate AI models to recommend optimal products for specific crops and soil types.
• [x] Validate API performance with real-world farming scenarios.

Nutrition Advice API: As a farmer, I want nutrition advice API endpoints for optimal crop growth. Tasks:

• [x] Create API endpoints to provide nutrient recommendations based on crop stages and soil health.
• [x] Implement AI/ML models for nutrient optimization.
• [x] Test API responses against known nutritional requirements.

Increased Crop Yield Insights API: As an agricultural researcher, I need API endpoints to analyze and predict factors contributing to increased crop yield. Tasks:

• [x] Develop API endpoints to analyze historical data and predict crop yield improvements.
• [x] Integrate AI/ML models for yield optimization recommendations.
• [x] Validate API accuracy with historical crop yield data.

Goals & Mid-Point Milestone

Mid-Point Goals:

API Development (Mid-Point Milestone):

• [x] Completion of API endpoints for personalized agri-advice, disease prediction, and seed/fertilizer recommendations.
• [x] Integration of AI/ML models with preliminary testing for accuracy and performance.
• [x] Initial Swagger documentation outlining API specifications.

Final Goals:

API Integration and Testing (Final Goal):

• [x] Full integration of AI/ML models with all API endpoints for agronomy features.
• [x] Comprehensive testing of APIs with real-world agricultural data to ensure accuracy and reliability.
• [x] Complete Swagger documentation for all API endpoints, including descriptions, parameters, example requests/responses, and usage guidelines.

Setup/Installation

No response

Expected Outcome

• A suite of robust APIs providing personalized agri-advice, disease prediction, seed/fertilizer recommendations, nutrition advice, and crop yield insights.
• Swagger documentation that facilitates easy integration and usage of the APIs by developers and agricultural stakeholders.

Acceptance Criteria

API Functionality:

APIs should accurately provide personalized advice, disease predictions, and recommendations based on AI/ML models. Responses should be timely, reliable, and culturally sensitive, considering all Indian regional languages.

Swagger Documentation:

Swagger documentation should comprehensively outline each API endpoint, including usage instructions and example scenarios.

Implementation Details

Go / Node.JS / Python (any one technology to be used)

Mockups/Wireframes

No response

Product Name

Agri Doctor

Organisation Name

Protean

Domain

⁠Agriculture

Tech Skills Needed

Swagger

Mentor(s)

@piyushmittaltech @RaviV-Protean @amitknamdeo @sunilnsdlekyc

Category

API

[Gitesg]

Hello @piyushmittaltech,

I'm eager to contribute to this project by developing a Flask backend that incorporates AI/ML models for data handling and recommendations.

[Nitish0777]

Hello @piyushmittaltech, @RaviV-Protean @amitknamdeo @sunilnsdlekyc "I'm eager to contribute to this project by developing a Backend in Node.js for an optimized API. Known for my problem-solving skills, I'm ready to tackle challenges and drive progress."

[saqib-masood]

Hello @piyushmittaltech I have already written API Docs for almost 60+ API's for a fintech startup in swagger . Apart from that I have also made multiple projects on AI/ML and also achieved Outstanding grade in ML in my coursework. I am interested in this project as well

[AbhimanyuSamagra]

Do not ask process related questions about how to apply and who to contact in the above ticket. The only questions allowed are about technical aspects of the project itself. If you want help with the process, you can refer instructions listed on Unstop and any further queries can be taken up on our Discord channel titled DMP queries. Here's a Video Tutorial on how to submit a proposal for a project.

[Anushkabh]

@piyushmittaltech @RaviV-Protean @amitknamdeo @sunilnsdlekyc
I am interested in writing proposal for this project .. are there any existing issues which i can solve and include in my proposal? i have one more question do we need to develop the models from scratch also? or just integrate the existing one?

[Harine19]

@piyushmittaltech @RaviV-Protean @amitknamdeo @sunilnsdlekyc I have implemented parts of this project and have some doubts: My APIs will accept and return responses in English for now, In the requirement I see that multilingual support is needed and I would like to know whether the API should handle all languages in input and output? Where exactly should the translation happen?

[jay05narayan]

@piyushmittaltech @RaviV-Protean @amitknamdeo @sunilnsdlekyc I am interested in contributing to this project by integrating agri related problems that faced by farmers like weather ,soil condition, season, market expectations of crops and solutions to it.Like an AI consultancy.

[Taeju1330]

Hello @piyushmittaltech I am very interested in contributing to this project as it aligns with my interest's. I would love to leverage my skills for its development.

[piyushmittaltech]

Weekly Goals

Week 1: Initial Subject Research and Problem Statement Definition

• [x] Discuss and define the problem statement in an initial meeting.
• [x] Identify target crops and related diseases.
• [x] Conduct research on crop diseases and respective nutrition.

Week 2: System Setup and R&D for Machine Learning Models

• [x] Install required software and tools.
• [x] Research crop disease cures and nutrition.
• [x] Identify development requirements (datasets, algorithms, APIs).
• [x] Search for relevant datasets for crops. (1000 images+)

Week 3: Data Cleaning and Dataset Creation

• [x] Extract data from datasets/augment own dataset.
• [x] Clean and pre-process the data.
• [x] Create/arrange high-quality datasets for the selected crops.
• [x] Research CNN models for crop and disease prediction.

Week 4: Dataset Finalization and Flowchart Creation

• [x] Finalize dataset for crop diseases.
• [x] Develop initial code for tomato disease detection.
• [x] Review weather API.

Week 5: Weather API and ML Model Development

• [x] Implement City Search API.
• [x] Use weather data to improve prediction accuracy.
• [x] Combine ML model output and weather data for nutrition report.

Week 6: Model Development and Final Presentation Preparation

• [x] Use Current Condition API to retrieve weather data.
• [x] Write code for crop and disease prediction.
• [x] Present progress at midpoint review.

Week 7: Crop Type and Disease Prediction Model

• [x] Integrate appropriate ML model with pooling and dense layers.
• [x] Apply image augmentation techniques for input data.
• [x] Train and save the model for API use.

Week 8: Model and API Integration

• [x] Deploy inference function for disease prediction.
• [x] Fine-tune the model.
• [x] Create Flask API with multiple endpoints.
• [x] Develop code to interact with city API.

Week 9: City Search and Current Weather API Integration

• [x] Extract location key from City Search API.
• [x] Process real time weather data for customized response.
• [x] Complete API data extraction.

Week 10: API Integration with ML Model

• [x] Mount saved crop and disease prediction model.
• [x] Pass new image to the model for prediction.
• [x] Integrate weather, city, crop, and disease details in a prompt.
• [x] Debug API using Postman.

Week 11: Customized Response

• [x] Set up response models (Gemma, LLaMA, Qwen).
• [x] Optimize response times with different models.
• [x] Implement Qwen for multilingual responses.

Week 12: Swagger Testing and Documentation

• [x] Create Swagger documentation for API endpoints.
• [x] Set up blueprint for Swagger.json.
• [x] Enable interactive API testing via Swagger UI.

[Harine19]

Weekly Learnings

Week 1: Initial Subject Research and Problem Statement Definition

• [x] Set up a communication channel and met with the mentor to discuss project details.
• [x] Studied the project description and requirements, focusing on increasing crop yield, with farmers as the target audience.
• [x] Defined specific problems to address, including disease prediction, cure advisory, and nutrition advice, from case studies in the issue ticket.
• [x] Identified target crops commonly grown in India, selecting wheat, maize, tomato, rice, and potato for project implementation.
• [x] Conducted research to identify the most prevalent diseases for each target crop.

Week 2: System Setup and R&D for Machine Learning Models

• [x] Installed necessary software and tools, including Python, Jupyter Notebook, and Git.
• [x] Conducted research on cures and nutrition required for identified crops based on the diseases, including pesticide, fertilizer, and NPK values.
• [x] Researched the development requirements, including image datasets, disease prediction ML algorithms, nutrition recommendation models, weather API, and Indic language translation API.
• [x] Searched for publicly available datasets related to crop diseases, nutrition, and agricultural practices from sources like Kaggle and government-backed databases.

Week 3: Data Cleaning and Dataset Creation

• [x] Identified and extracted data from suitable datasets for the five selected crops.
• [x] Cleaned and pre-processed the data to handle inconsistencies and duplicates.
• [x] Created high-quality datasets with 500-1,000 images for each crop.
• [x] Researched existing CNN machine learning models for disease prediction based on image classification for each of the five crops.

Week 4: Dataset Finalization and Flowchart Creation

• [x] Uploaded the collected datasets to a shared drive for easy access.
• [x] Developed a flowchart representing the disease detection and nutrition advice process.
• [x] Created detailed flowcharts for each crop, outlining disease identification and nutrition recommendation.
• [x] Developed an initial AI/ML code snippet for disease detection in tomatoes.
• [x] Reviewed the weather API provided by the mentor and studied its functionality.

Week 5: Weather API and ML Model Development

• [x] Used the AccuWeather API and assessed the JSON output using an API key.
• [x] Determined how weather conditions affect crop yield and nutrition absorption, factoring these into predictions.
• [x] Implemented the ‘City Search’ API to obtain location details from farmer inputs.
• [x] Researched how to combine the weather API data and ML model output to generate comprehensive nutrition reports.

Week 6: Model Development and Final Presentation Preparation

• [x] Used the ‘Current Condition’ API to retrieve weather conditions based on location keys.
• [x] Wrote code to predict crop types and diseases using a CNN.
• [x] Created a PowerPoint presentation for the midpoint review by C4GT reviewers, covering the project’s progress.
• [x] Prepared a speech and demonstration for the presentation.

Week 7: Crop Type and Disease Prediction Model

• [x] Integrated the ResNet50 pre-trained model with global average pooling and dense layers.
• [x] Applied image augmentation techniques to enhance training data and model generalization.
• [x] Compiled the model using the Adam optimizer and trained it to classify diseases in five crops (corn, potato, rice, tomato, and wheat).
• [x] Packaged and saved the trained model for API integration.

Week 8: Model and API Integration

• [x] Deployed an inference function to process new images and predict diseases from 20 plant disease classes.
• [x] Fine-tuned the model to improve performance and generalization.
• [x] Created a Flask API backend with multiple endpoints for city weather, current weather conditions, and plant disease detection.
• [x] Developed code to interact with the city API and retrieve JSON data.

Week 9: City Search and Current Weather API Integration

• [x] Processed the JSON output to extract the location key.
• [x] Passed the location key to the weather API to retrieve weather data.
• [x] Processed the JSON output from the weather API to generate customized responses.
• [x] Extracted key parameters like temperature for the final response.
• [x] Completed all required API data extraction.

Week 10: API Integration with ML Model

• [x] Mounted the saved crop and disease prediction model.
• [x] Passed a new image to the model and processed its output.
• [x] Finalized code to integrate weather, city, crop, and disease details into a prompt.
• [x] Used Postman to debug and test API integrations.
• [x] Installed and set up Ollama for prompt processing.

Week 11: Customized Response

• [x] Researched models like Gemma, LLaMA, and Qwen to improve responses.
• [x] Developed code to serve the Ollama model and experimented with different models to optimize response times.
• [x] Used Qwen2:1.5b to generate multilingual responses in Hindi, Bengali, Urdu, and more.

Week 12: Swagger Testing and Documentation

• [x] Configured Swagger UI for API testing and documentation.
• [x] Created a detailed Swagger specification (swagger.json) to describe endpoints, response structures, and input parameters.
• [x] Set up a blueprint and linked it to the swagger.json file.
• [x] Enabled easy API testing through Swagger UI, allowing users to visualize and test endpoints directly.