mivia 0.1.3

Python API client for MiViA (Microstructure Analysis)

MiViA Python Client

Python API client for MiViA (Microstructure Analysis).

Installation

pip install mivia

Configuration

Set your API key as environment variable:

For linux:

export MIVIA_API_KEY="your-api-key"
export MIVIA_BASE_URL="https://app.mivia.ai/api"  # optional, default
export MIVIA_PROXY="http://proxy:8080"  # optional, proxy URL

For Windows (Powershell):

$env:MIVIA_API_KEY = 'your-api-key'
$env:MIVIA_BASE_URL = 'https://app.mivia.ai/api'  # optional, default
$env:MIVIA_PROXY = 'http://proxy:8080'  # optional, proxy URL

Usage

Async Client

import asyncio
from mivia import MiviaClient

async def main():
    async with MiviaClient() as client:
        # List models
        models = await client.list_models()
        print(f"Models: {[m.display_name for m in models]}")

        # High-level: analyze images
        jobs = await client.analyze(
            file_paths=["image1.png", "image2.png"],
            model_id=models[0].id,
            wait=True,
        )

        for job in jobs:
            print(f"Job {job.id}: {job.status}")

        # Download report
        await client.download_pdf(
            job_ids=[j.id for j in jobs],
            output_path="report.pdf",
        )

asyncio.run(main())

Sync Client

from mivia import SyncMiviaClient

client = SyncMiviaClient()

# List models
models = client.list_models()

# Get customizations for a model
customizations = client.get_model_customizations(models[0].id)

# Upload and analyze with customization
jobs = client.analyze(
    file_paths=["image.png"],
    model_id=models[0].id,
    customization_id=customizations[0].id if customizations else None,
)

# Download report
client.download_csv(
    job_ids=[j.id for j in jobs],
    output_path="report.zip",
)

CLI

# List models
mivia models

# List customizations for a model
mivia customizations MODEL_UUID

# Upload images
mivia upload image1.png image2.png

# Analyze (upload + run + wait) - supports model name or UUID
mivia analyze image.png --model "Decarburization"
mivia analyze image.png --model MODEL_UUID

# Analyze with customization (by name or UUID)
mivia analyze image.png --model "Test" --customization "Template Name"

# List available customizations for analyze
mivia analyze --model MODEL_UUID --list-customizations

# List jobs
mivia jobs list

# Get job details
mivia jobs get JOB_UUID

# Wait for jobs
mivia jobs wait JOB_UUID1 JOB_UUID2

# Download PDF report
mivia report pdf JOB_UUID -o report.pdf

# Download CSV report
mivia report csv JOB_UUID -o report.zip --no-images

# Show config
mivia config

# Use proxy
mivia --proxy http://proxy:8080 models

Proxy Configuration

Proxy can be configured via parameter, CLI option, or environment variable.

Library

from mivia import MiviaClient, SyncMiviaClient

# Via parameter (highest priority)
async with MiviaClient(proxy="http://proxy:8080") as client:
    models = await client.list_models()

# Sync client
client = SyncMiviaClient(proxy="http://proxy:8080")
models = client.list_models()

# Via environment variable (MIVIA_PROXY)
# export MIVIA_PROXY="http://proxy:8080"
async with MiviaClient() as client:
    models = await client.list_models()

CLI

# Via --proxy option
mivia --proxy http://proxy:8080 models
mivia --proxy http://proxy:8080 analyze image.png --model "Decarburization"

# Via environment variable
export MIVIA_PROXY="http://proxy:8080"
mivia models

# Verify proxy configuration
mivia --proxy http://proxy:8080 config

Priority Order

1. Explicit proxy parameter / CLI --proxy option
2. MIVIA_PROXY environment variable
3. Standard proxy env vars (HTTP_PROXY, HTTPS_PROXY) via httpx defaults
4. No proxy (direct connection)

API Reference

MiviaClient

Method	Description
upload_image(path)	Upload single image
upload_images(paths)	Upload multiple images
list_images()	List uploaded images
delete_image(id)	Delete image
list_models()	List available models
get_model_customizations(id)	Get model customizations
create_jobs(image_ids, model_id)	Create computation jobs
get_job(id)	Get job details with results
list_jobs()	List jobs with pagination
wait_for_job(id)	Poll until job completes
wait_for_jobs(ids)	Wait for multiple jobs
download_pdf(job_ids, path)	Download PDF report
download_csv(job_ids, path)	Download CSV report
analyze(paths, model_id)	High-level: upload + run + wait