Here is a link to a script that shows how to extracts SSNs from a page text: Acrobat Javascript Samples Scripts. In this case, all words need to be enumerated using this.getPageNthWord() method and the resulting text string can be searched using a regular expression. Then there is no need to measure text coordinates and text can be extracted by performing a text search. The similar script can be used if the text of interest conforms to a well-defined text pattern (such as SSN, account number, client ID, email and etc.). Can be a very slow process if there is a lot of text and a lot of files involved. If new documents with a different layout are used, then the script needs to be edited with new coordinates.ģ. Coordinates for the text needs to be measured manually.Ģ.
Once done enumerating the words, save the file using this.SaveAs() method while combining a predefined output path with the extracted text.ġ.
If the word lies inside an "area of interest", add it to the output text string.ĥ. Quads are returned in the page coordinate system (user space, origin is in bottom-left corner), while coordinates on screen are measured in screen coordinate system (device space, origin is in upper-left corner, y-axis is pointing down).Ĥ. Complication: it is necessary to take page rotation and difference in coordinate systems into account when checking each quad.
Check each "quad" returned by the function against the area(s) of interest (determined manually in step 1). In JavaScript code, enumerate all words on the page using this.getPageNthWordQuads() method that returns coordinates for each word as an array of coordinates.ģ. If there are multiple areas, measure each one and record the coordinates.Ģ.
First, use a ruler/grid tool (Use Ctrl+U and Ctrl+R keyboard shortcuts to toggle the tools on/off) to determine the coordinates on the page where the "text of interest" is located. If it is necessary to extract text from a specific location on a page in order to use it in the output file name, then the algorithm is as follows:ġ.
0 kommentar(er)
