Done
Updated curl, exiftool, Java, jq, Python, Tesseract-OCR and even more important, finally switched from sequential calls with tika-app for file content extraction to using tika-pipes! Result is an enormous performance boost for content extraction because object stays alive and multiple clients and emitters are used in parallel. Processing time for the default test data set went down from 350+ to 8 seconds! 😱
INFO [pool-2-thread-9] 01:23:35,996 org.apache.tika.pipes.PipesClient pipesClientId=7: starting process
INFO [pool-2-thread-2] 01:23:35,996 org.apache.tika.pipes.PipesClient pipesClientId=0: starting process...
INFO [pool-2-thread-6] 01:23:36,001 org.apache.tika.pipes.PipesClient pipesClientId=3: starting process
INFO [main] 01:23:44,087 org.apache.tika.async.cli.TikaAsyncCLI Successfully finished processing 45 files in 8219 ms
Next
Further testing of pipes exploring optimal parameters settings and running on larger datasets.
