Change Details

Create a wrapper API on Tool Labs to interact with the Google Vision API to perform OCR on images from WikiSource. It will need to use [[ https://wikitech.wikimedia.org/wiki/Nova_Resource:Google-api-proxy/Documentation | Bryan's Google API proxy ]] since Google's APIs require IP address whitelisting (and Tool Labs sends requests from random nodes with different IP addresses). Acceptance criteria: * The wrapper API should take the following input: language code, image URL. * It should accept requests only from wikisource.org. * It should then encode the image (base64 byte stream?), add the API credentials, set the language in the languageHints array, and submit the request as JSON to the API proxy. * It should return the OCRed text or an error message if there was an error. See Google Vision API documentation at https://cloud.google.com/vision/docs/ and Discovery URL at https://vision.googleapis.com/$discovery/rest?version=v1. See T140037#2507218 and T140037#2528369 for the format of the JSON data to submit. Ask @kaldari for the API key.

Create a wrapper API on Tool Labs to interact with the Google Vision API to perform OCR on images from WikiSource. It will need to use [[ https://wikitech.wikimedia.org/wiki/Nova_Resource:Google-api-proxy/Documentation | Bryan's Google API proxy ]] since Google's APIs require IP address whitelisting (and Tool Labs sends requests from random nodes with different IP addresses). Acceptance criteria: * The wrapper API should take the following input: language code, image URL. * It should accept requests only from wikisource.org. * Initially it should be limited to handling images that are no larger than 10 MB. * It should then encode the image (base64 byte stream?), add the API credentials, set the language in the languageHints array, and submit the request as JSON to the API proxy. * It should return the OCRed text or an error message if there was an error. See Google Vision API documentation at https://cloud.google.com/vision/docs/ and Discovery URL at https://vision.googleapis.com/$discovery/rest?version=v1. See T140037#2507218 and T140037#2528369 for the format of the JSON data to submit. Ask @kaldari for the API key.

Create a wrapper API on Tool Labs to interact with the Google Vision API to perform OCR on images from WikiSource. It will need to use [[ https://wikitech.wikimedia.org/wiki/Nova_Resource:Google-api-proxy/Documentation | Bryan's Google API proxy ]] since Google's APIs require IP address whitelisting (and Tool Labs sends requests from random nodes with different IP addresses). Acceptance criteria: * The wrapper API should take the following input: language code, image URL. * It should accept requests only from wikisource.org. * Initially it should be limited to handling images that are no larger than 10 MB. * It should then encode the image (base64 byte stream?), add the API credentials, set the language in the languageHints array, and submit the request as JSON to the API proxy. * It should return the OCRed text or an error message if there was an error. See Google Vision API documentation at https://cloud.google.com/vision/docs/ and Discovery URL at https://vision.googleapis.com/$discovery/rest?version=v1. See T140037#2507218 and T140037#2528369 for the format of the JSON data to submit. Ask @kaldari for the API key.