| 1 | ------------------------------------------------------------------------
|
|---|
| 2 | README FOR Greenstone USERS: TO SUPPORT ADDITIONAL LANGUAGES FOR OCR
|
|---|
| 3 | ------------------------------------------------------------------------
|
|---|
| 4 | Greenstone can be configured to use Tesseract to OCR images, and use Tika
|
|---|
| 5 | in combination with Tesseract to OCR PDFs.
|
|---|
| 6 |
|
|---|
| 7 | By default, the Greenstone Tesseract extension only comes with support for OCR-ing
|
|---|
| 8 | English and Onscreen Display text, as otherwise the extension will become too large.
|
|---|
| 9 |
|
|---|
| 10 | Tesseract supports OCR for many languages (for the scripts of many languages).
|
|---|
| 11 | The supported languages are at https://github.com/tesseract-ocr/tessdata,
|
|---|
| 12 | where they're indicated by their official 3 letter language code.
|
|---|
| 13 | (You can Google to find the 3 letter lang code for your languages of interest).
|
|---|
| 14 |
|
|---|
| 15 | To obtain support for other languages, you can do one of:
|
|---|
| 16 |
|
|---|
| 17 | a. manually download the <3-letter-langcode>.traineddata files for languages
|
|---|
| 18 | you want from https://github.com/tesseract-ocr/tessdata
|
|---|
| 19 |
|
|---|
| 20 | b. Run the following from the toplevel of your GS3 installation:
|
|---|
| 21 | source ./gs3-setup.sh
|
|---|
| 22 | cd gs2build/ext/tesseract/linux/share/tessdata
|
|---|
| 23 |
|
|---|
| 24 | Then for each language code, run the following with <3-letter-langcode> adjusted
|
|---|
| 25 | accordingly:
|
|---|
| 26 | wget https://github.com/tesseract-ocr/tessdata/raw/master/<3-letter-langcode>.traineddata
|
|---|
| 27 |
|
|---|
| 28 | c. You can download all the supported languages in one step if you have git
|
|---|
| 29 | installed. First move (or remove) the existing "tessdata" folder, before running
|
|---|
| 30 | git clone to get all the languages that have OCR support:
|
|---|
| 31 |
|
|---|
| 32 | cd gs2build/ext/tesseract/linux/share
|
|---|
| 33 | #rm -rf tessdata
|
|---|
| 34 | mv tessdata tessdata.basic
|
|---|
| 35 | git clone https://github.com/tesseract-ocr/tessdata
|
|---|
| 36 |
|
|---|
| 37 |
|
|---|
| 38 | ------------------------------
|
|---|
| 39 | Background Information:
|
|---|
| 40 | ------------------------------
|
|---|
| 41 | Greenstone can only index text in documents that contain extractable text.
|
|---|
| 42 | Not documents that only have images of text ("photos" of text don't contain
|
|---|
| 43 | selectable text).
|
|---|
| 44 |
|
|---|
| 45 | There is a process called OCR (Optical Character Recognition) to recognise
|
|---|
| 46 | any individual characters constituting text represented in images, and thereby
|
|---|
| 47 | produce the text in images that otherwise have no extractable text.
|
|---|
| 48 |
|
|---|
| 49 | Tesseract is OCR software licensed under the Apache 2.0 License. Tesseract can
|
|---|
| 50 | be used by Greenstone for OCR-ing images, to thus get text from those images
|
|---|
| 51 | which Greenstone can then index for full text searching on that image document.
|
|---|
| 52 |
|
|---|
| 53 | Tesseract cannot OCR PDFs, only images. However, Apache Tika can work with Tesseract
|
|---|
| 54 | (both licensed under the Apache 2.0 License) to OCR PDFs that contain pages
|
|---|
| 55 | which are only images of text rather than actual extractable text.
|
|---|
| 56 |
|
|---|
| 57 | Greenstone can use the combination of Apache Tika and Tesseract to further process
|
|---|
| 58 | any PDFs of images of text too, the OCR process producing text that Greenstone can
|
|---|
| 59 | index to enable full text searching on the original document (which otherwise
|
|---|
| 60 | contained no extractable text, only images of text).
|
|---|
| 61 |
|
|---|
| 62 | Important Notes:
|
|---|
| 63 |
|
|---|
| 64 | a. Where OCR is involved in any process, the quality of the OCR-ed text that is
|
|---|
| 65 | produced is tightly dependent on the quality of image files that went into the
|
|---|
| 66 | process. The higher the DPI (dots per inch) of the images and the clearer the
|
|---|
| 67 | legibility of the images of text that go into the digital OCR-ing process, the more
|
|---|
| 68 | sensible and accurate the OCR-ed text that results. In cases of poor quality images,
|
|---|
| 69 | gibberish will be produced. With average-quality input images, the OCR-ed text is a
|
|---|
| 70 | combination of text accurate to the original interspersed occasionally by strange
|
|---|
| 71 | characters.
|
|---|
| 72 |
|
|---|
| 73 | b. OCR is for recognising characters constituting text in images. Characters are
|
|---|
| 74 | components of scripts, and there are many language scripts in the world. As a result,
|
|---|
| 75 | in order for OCR to recognise the characters that constitute the script of the
|
|---|
| 76 | language your document contains, there needs to be support for that language's script
|
|---|
| 77 | in the OCR software used, in this case Tesseract.
|
|---|
| 78 |
|
|---|
| 79 | The languages' scripts that Tesseract supports (indicated by their 3 letter language
|
|---|
| 80 | codes) are at https://github.com/tesseract-ocr/tessdata
|
|---|
| 81 |
|
|---|
| 82 | By default, the Greenstone Tesseract extension only comes with support for OCR-ing
|
|---|
| 83 | English and Onscreen Display text, as otherwise the extension will become too large.
|
|---|
| 84 |
|
|---|
| 85 | To allow the Greenstone Tesseract extension to OCR further languages that
|
|---|
| 86 | Tesseract already supports, read the section "TO SUPPORT ADDITIONAL LANGUAGES FOR OCR".
|
|---|
| 87 |
|
|---|
| 88 | ------------------------------------------------------------------------
|
|---|
