Skip to content

data-structure.md

Latest commit

 

History

History
71 lines (59 loc) · 2.31 KB

data-structure.md

File metadata and controls

71 lines (59 loc) · 2.31 KB

Data Structure Guide

Data Model

Plant Identification

The system uses a two-level identification system:

  • plant_name: Identifies individual plants (e.g., "plant1", "plant2")
  • gene_variety: Links plants to their genetic variety (e.g., "11430")

Data Components

Each plant entry contains:

  • Plant Name and Gene Variety
  • Sensor Data (temperature, humidity, light)
  • Images (stored in GridFS, referenced by image_id)
  • Genomic Data (linked via Gene Variety)

Directory Structure

 data/
 ├── plant1/
 │ ├── images/
 │ │ ├── image1.png
 │ │ └── ...
 │ └── sensor_data.csv
 ├── plant2/
 │ ├── images/
 │ │ ├── image1.png
 │ │ └── ...
 │ └── sensor_data.csv
 └── genomic_data.json

Data Organization

  • Each plant has its own directory (plant1, plant2, etc.)
  • Plant images are stored in the images subdirectory
  • Sensor data is stored in sensor_data.csv within each plant directory
  • Genomic data is stored in the root of the data directory

Image Storage

Images are stored using MongoDB's GridFS system:

  • Images are split into chunks for efficient storage
  • Each image is referenced by an image_id in the plant data
  • Images can be accessed through MongoDB Express or the MongoDB shell

Accessing Images

Through MongoDB Express:

  1. Navigate to http://plant-data.local/mongo-express
  2. Go to the plant_data database
  3. Look for the fs.files and fs.chunks collections
  4. Images can be found in fs.files with their metadata
  5. The actual image data is stored in fs.chunks

Through MongoDB Shell:

# Connect to MongoDB
kubectl exec -it $(kubectl get pod -l app=mongodb -o jsonpath='{.items[0].metadata.name}') -- mongosh

# Switch to plant_data database
use plant_data

# Find image metadata by ID
db.fs.files.find({"_id": ObjectId("YOUR_IMAGE_ID")})

# Export image (from your local machine)
kubectl exec -it $(kubectl get pod -l app=mongodb -o jsonpath='{.items[0].metadata.name}') -- mongofiles --db=plant_data get_id 'YOUR_IMAGE_ID' --local=downloaded_image.jpg

Mounting Configuration

  • The data directory is mounted into the kind cluster at /data
  • The plant-sender pod accesses data through /app/data
  • Directory structure must match the expected format