v5.0.0.alpha1

¡Breaking Changes!

• 💥 ♻️ update V1 & V2 syntaxes to match other SDKs
  • ♻️ move V1 client to V1 module
  • ♻️ move V2 client to V2 module
  • ♻️ move legacy products to 'V1' module
  • ♻️ add parsing and extraction to V1 module
  • ♻️ move V1 Http, error, image, pdf to V1 namespace
  • ♻️ move V2 Http, error, image, pdf to V2 namespace
  • ♻️ move data schema to extraction parameters namespace
• 💥 ♻️ remove /v1 and /v2 from base URLs to match other SDKs
• 💥 ♻️ remove deprecated methods and classes
• 💥 🔧 move LocalInputSource page_count to an attribute
• 💥 ♻️ Remove V2 Inference in favor of extraction
• 💥 ♻️ Move PDF & Image operations to their respective namespaces and remove Extraction namespace
• 💥 ♻️ Harmonize all acronyms as proper camelCase & PascalCase
• 💥 ⬆️ bump minimal supported version to python 3.10
• 💥 ♻️ remove sourceFrom initializers in favor of direct InputSource initialization
• 💥 remove support for the following V1 products:
  • ⚰️ Bill of Lading V1
  • ⚰️ Business Card V1
  • ⚰️ Delivery Notes V1
  • ⚰️ Driver License V1
  • ⚰️ (FR) Energy Bill V1
  • ⚰️ (FR) Healthcare Card V1
  • ⚰️ (IND) Passeport V1
  • ⚰️ Material Certificate V1
  • ⚰️ Nutrition Fact V1
  • ⚰️ (FR) Payslip V3
  • ⚰️ Resume V1
  • ⚰️ (US) Healthcare Cards V1
  • ⚰️ (US) Mail V3
• 💥 ♻️ switch to httpx from requests
• 💥 ✨ add support for mindee-lite package without heavier dependencies

Changes

• ✨ add support for model search from the V2 client
• ✨ add crop and split extractors from product classes
• ✨ add support for direct url GET inference
• ✨ add support for custom HTTPx client
• 🔧 add support for better tooling
• ⬇️ remove pytz dependency
• ♻️ move V2 API-specific internals to their own file

Fixes

• 🐛 fix some circular import shenanigans
• 🐛 fix memory leak issues when altering PDF pages

Full Changelog: v4.36.1...v5.0.0.alpha1