summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2019-12-13Hopefully fix empty training bugNick White
2019-12-13Mention training in ocr error messageNick White
2019-12-13Print stdout and stderr output when tesseract failsNick White
2019-12-11Add addtoanalysequeue tool, which is useful for debuggingNick White
2019-12-11Fix typo incorrectly screwing up PDFsNick White
2019-12-11Clarify use of -training in toolsNick White
2019-12-11Clean up and correct book name parsing in the pipeline, and update usage of ↵Nick White
getpipelinebook
2019-12-11Add ability to set a different training for the ocr jobNick White
2019-12-11Use aws.go with mkpipeline too, plus fix one log.Fatal call in aws.go which ↵Nick White
should have been handled by caller
2019-12-06Don't abort PDF generation if pages aren't found, just do the best that can ↵Nick White
be done and move on; not all books will have all page types (such as wipeonly books)
2019-12-05Remove (the generally empty) files in the case of a failed downloadNick White
2019-12-05Default getpipelinebook to downloading pdfs instead of imagesNick White
2019-12-05Fix the PDF in analyse step part of bookpipelineNick White
2019-12-05Add pdf generation to analyse step (untested)Nick White
2019-12-03Rewrite lspipeline book listing part to be much faster by taking advantage ↵Nick White
of the aws CommonPrefixes output
2019-12-03Don't pause between OCR page jobs; this should save us significant amounts ↵Nick White
of time when there are large numbers of pages
2019-11-29Make error message clear what page is causing issuesNick White
2019-11-26Improve usage noticeNick White
2019-11-26Ensure error in file walking is correctly returnedNick White
2019-11-20Add x/image to go.modNick White
2019-11-20Merge branch 'addpdf'Nick White
2019-11-20Implement image resizing option into PDF generation, so that smaller PDFs to ↵Nick White
be generated
2019-11-19Send pages to the individual OCR Page queue by defaultNick White
This now concludes the OCR Page queue stuff; it should all be working out of the box now.
2019-11-19Add ocrpage queue for processing individual pagesNick White
This should be a good way to get around the ongoing heartbeat issue, as individual page jobs will never come close to a the 12 hour mark that can cause the bug. The OCR page processing is done and working now, still to do is to populate the queue (rather than the ocr queue) after preprocessing / wiping.
2019-11-12Merge branch 'addpdf'Nick White
2019-11-12Embed a font, compressed, into the binaryNick White
2019-11-12Fix sleep in unstickocrNick White
2019-11-12Add unstickocr tool, until the heartbeat bug is eliminatedNick White
2019-11-12Add spotme command to start appropriate spot instancesNick White
2019-11-12Merge branch 'addpdf'Nick White
2019-11-11Add go.mod and go.sumNick White
2019-11-11Switch to main gofpdf, now our SetTextRenderingMode has been mergedNick White
2019-11-01Compress the font with zlib, and include it in repoNick White
2019-10-31Add capability to embed font files into toolNick White
2019-10-31PDF: add functionality to use "best" file if it existsNick White
2019-10-31PDF: add space to each word to ensure copy-past ability from more PDF readersNick White
2019-10-31PDF: lay out every word with coordinates separatelyNick White
I presumed this would mean that multiple words next to each other couldn't be reliably searched for, but this seems not to be the case.
2019-10-31Add flag to switch between binarised and colour outputNick White
2019-10-31Move PDF handling code to a separate fileNick White
2019-10-31Many improvements to pdfbook; basically working nowNick White
2019-10-31Add work in progress PDF producerNick White
2019-10-29Print heartbeat error on failureNick White
2019-10-29Debugging: kill process immediately a heartbeat error is detected (systemd ↵Nick White
will restart it soon thereafter)
2019-10-29Another attempt to fix the ongoing heartbeat issueNick White
This time wait up to 1 second between attempts, reduce long polling time significantly, and attempt for longer before giving up.
2019-10-28Try to fix heartbeat renew issue more fullyNick White
This approach first sets the remaining visibility timeout to zero. This should ensure that the message is available to re-find as soon as the process looks for it. Correspondingly the delay between checks is much shorter, as there shouldn't be a reason for much delay.
2019-10-23getpipelinebook: default to downloading corresponding page images, and add ↵Nick White
option to download the original page images too
2019-10-23Manually calculate yticks, so they fall on reasonable numbersNick White
2019-10-23Add more annotations to graph; anything outside of the 80% "normal" band ↵Nick White
gets an annotation now, and that band is labelled
2019-10-17Adjust the heartbeat searching function to hopefully have better luck at ↵Nick White
finding it and not letting another process steal it.
2019-10-16Rewrite booktopipeline to use bookpipeline aws interfaceNick White