page.add_highlight_annot(start=pointa, stop=pointb) not working

[kalimer00]

Hello,

i'm trying to highlight a text in a pdf from a start word "pointa" to an end word "pointb" but it wont work it will mark all the text on the page Maybe some one could help me (pleas) and figure out what i'm doing wrong.

output:

`import fitz
import config
class FindTextblock:

    doc = fitz.open(config.fpath_int+"/test_source.pdf")
    for page in doc:
        wordlist = page.get_text("words")
        wordlist.sort(key=lambda w: (w[1], w[0]))  # sort vertical, then horizontal
        pointa = None
        pointb = None
        for w in wordlist:
            if w[4] == "Therapievorschlag":  # choose a word to start marking
                pointa = fitz.Point(w[:2])  # top left of work rectangle
                break
        for w in wordlist:
            if w[4] == "EU-Verordnung":  # some word for stopping the marking
                pointb = fitz.Point(w[2:4])  # bottom right of word rectangle
                break

        page.add_highlight_annot(start=pointa, stop=pointb)

    doc.save("test.pdf")

FindTextblock()`

test file:
test_source.pdf

Thank you

[JorjMcKie]

@kalimer00 - welcome and hello again!
Thanks for using this channel. I will have a look.

[JorjMcKie]

Made minor adjustments to your script, now it does work perfectly:

The modified script:

import fitz

doc = fitz.open("test_source.pdf")
for page in doc:
    wordlist = page.get_text("words", sort=True)  # has built-in parameter
    pointa = None
    pointb = None
    for w in wordlist:
        if w[4] == "Therapievorschlag":  # choose a word to start marking
            pointa = fitz.Point(w[:2])  # top left of work rectangle
            break
    for w in wordlist:
        if w[4] == "EU-Verordnung":  # some word for stopping the marking
            pointb = fitz.Point(w[2:4])  # bottom right of word rectangle
            break

    if pointa != None and pointb != None:  # guard against hick ups
        page.add_highlight_annot(start=pointa, stop=pointb)

doc.save("test.pdf")

[JorjMcKie]

This script is even a bit simpler:

import fitz

doc = fitz.open("test_source.pdf")
for page in doc:
    rl1 = page.search_for("Therapievorschlag")
    if rl1 == []:  # not on page
        continue
    rl2 = page.search_for("EU-Verordnung")
    if rl2 == []:  # not on page
        continue
    pointa = rl1[0].tl
    pointb = rl2[0].br
    page.add_highlight_annot(start=pointa, stop=pointb)

doc.save("test.pdf")

[kalimer00]

wow, that was fast. thank you so much, it works like a charm.

[JorjMcKie]

always welcome!

[kalimer00]

hello, the code works fine when im execute it on a single file, but if i loop through a folder with pdf files it will damage the output files and when i try to open them they will throw the error massage “A number is out of range”

`import fitz, os, config

def mark_text_block(path, start_word, end_word):

number = 0
# doc = fitz.open(config.fpath_int+"/test_source.pdf")
res = os.listdir(path)
for file in res:
    if file.endswith(".pdf"):
        print("File in folder: "+file)
        print("Path to file: " +path+file)
        print("-------------------------")
        doc = fitz.open(path+file, filetype="pdf")
        for page in doc:
            wordlist = page.get_text("words", sort=True)  # has built-in parameter
            pointa = None
            pointb = None
            for w in wordlist:
                if w[4] == start_word:  # choose a word to start marking
                    pointa = fitz.Point(w[:2])  # top left of work rectangle
                    break
            for w in wordlist:
                if w[4] == end_word:  # some word for stopping the marking
                    pointb = fitz.Point(w[2:4])  # bottom right of word rectangle
                    break
            if pointa != None and pointb != None:  # guard against hick ups
                page.add_highlight_annot(start=pointa, stop=pointb)

        number = number + 1
        doc.save(str(number)+"_"+file)

mark_text_block(config.fpath_int, config.start_text_2, config.end_text)`

[JorjMcKie]

There is no way to help without reproducible, sorry. Maybe to clarify a few questions:

• Can you reproduce with a directory with only one file?
• Does the same file work in standalone, but no longer when in a directory?
• How do to protect against multiple changes of the same file? If you prefix the filename with a number, this may happen in a second go, right? Try to write in a separate directory, or choose a naming pattern for new files such that they can be safely skipped. E.g. doc.save(doc.name.replace(".pdf", "-marked.pdf")) and skip files ending like this on input.

[JorjMcKie]

Oh please do so!
There always is a chance to improve ...

[JorjMcKie]

Received your example PDFs and rechecked, thank you. Everything works fine - could not detect any problem.

[kalimer00]

You could apply the annotation on both PDF files without the error massage "a number is out of range" after opening?

[JorjMcKie]

No! Instead I found a problem: because of how this PDF is created, corners of infinite text rectangles are stored in the annotation definition. Those are the number Adobe Acrobat is complaining about.
Other PDF viewer are more tolerant and simply ignore obvious nonsense.
Will be fixed in the next version.

[kalimer00]

Aaaah ok perfekt, thanks for your help