pip install a-pandas-ex-tesseract-multirow-regex-fuzzfrom a_pandas_ex_tesseract_multirow_regex_fuzz import pd_add_regex_fuzz_multiline,pd_add_tesseract
pd_add_tesseract(tesseractpath=r"C:\Program Files\Tesseract-OCR\tesseract.exe")
pd_add_regex_fuzz_multiline()
df=pd.Q_Tesseract_to_DF(r'https://i.ytimg.com/vi/fa82Qpw6lyE/hqdefault.jpg') level page_num block_num par_num ... width height conf text
0 1 1 0 0 ... 480 360 -1
1 2 1 1 0 ... 275 177 -1
2 3 1 1 1 ... 258 109 -1
3 4 1 1 1 ... 222 19 -1
4 5 1 1 1 ... 23 14 95.939438 No
5 5 1 1 1 ... 33 16 96.663704 stop
6 5 1 1 1 ... 42 18 93.283119 signs,
7 5 1 1 1 ... 65 19 82.424248 speedin'
8 5 1 1 1 ... 37 15 96.098083 limit
9 4 1 1 1 ... 245 19 -1
10 5 1 1 1 ... 74 19 79.710175 Nobody's
11 5 1 1 1 ... 49 14 95.714142 gonna
12 5 1 1 1 ... 37 14 96.856789 slow
13 5 1 1 1 ... 24 9 95.754181 me
14 5 1 1 1 ... 43 14 96.090439 down
15 4 1 1 1 ... 209 19 -1
16 5 1 1 1 ... 32 15 96.436874 Like
17 5 1 1 1 ... 8 10 96.436874 a
18 5 1 1 1 ... 49 18 94.924347 wheel,
19 5 1 1 1 ... 49 14 92.669563 gonna
20 5 1 1 1 ... 32 18 96.724014 spin
21 5 1 1 1 ... 12 14 95.647652 itFuzzsearch in multiple rows in a DataFrame - works with any DataFrame, not only tesseract DataFrames
fuzzdfsearch = df.ds_fuzz_multirow(column='text',fuzzsearch='Rocking BAND')
level page_num block_num par_num line_num word_num left top width height conf text aa_weighted aa_len aa_npsum aa_weight aa_whole_text old_index aa_whole_text_len aa_whole_text_len_difference
aa_npsum aa_weight aa_whole_text aa_whole_text_len aa_whole_text_len_difference
1 180.000000 rockin' band 12 0 5 1 1 2 1 4 128 194 55 14 86.868996 rockin' 90.000000 7 1 180.000000 rockin' band 0 12 0
0 5 1 1 2 1 5 188 194 40 14 96.840813 band 90.000000 4 1 180.000000 rockin' band 1 12 0
130.000000 Playin’ in 10 2 5 1 1 2 1 1 37 194 53 19 41.302063 Playin’ 40.000000 7 1 130.000000 Playin’ in 12 10 2
2 5 1 1 2 1 2 95 194 14 14 95.126900 in 90.000000 2 1 130.000000 Playin’ in 13 10 2
122.142857 promise land 12 0 5 1 1 4 1 7 207 239 64 19 95.758591 promise 45.000000 7 1 122.142857 promise land 22 12 0
0 5 1 1 4 1 8 276 239 36 14 89.825577 land 77.142857 4 1 122.142857 promise land 23 12 0
96.428571 Satan! Paid 11 1 5 1 1 1 5 2 74 172 49 14 96.457039 Satan! 45.000000 6 1 96.428571 Satan! Paid 56 11 1
1 5 1 1 1 5 3 128 171 34 15 96.457039 Paid 51.428571 4 1 96.428571 Satan! Paid 57 11 1
93.857143 spin it 7 5 5 1 1 1 3 5 197 127 32 18 96.724014 spin 48.857143 4 1 93.857143 spin it 60 7 5
5 5 1 1 1 3 6 234 127 12 14 95.647652 it 45.000000 2 1 93.857143 spin it 61 7 5
90.000000 Look at 7 5 5 1 1 3 1 3 134 216 38 15 96.986824 Look 45.000000 4 1 90.000000 Look at 28 7 5
5 5 1 1 3 1 4 177 219 15 12 95.994835 at 45.000000 2 1 90.000000 Look at 29 7 5
84.428571 Like a 6 6 5 1 1 1 3 1 37 126 32 15 96.436874 Like 24.428571 4 1 84.428571 Like a 24 6 6
6 5 1 1 1 3 2 74 131 8 10 96.436874 a 60.000000 1 1 84.428571 Like a 25 6 6
79.200000 way to 6 6 5 1 1 4 1 4 123 244 30 14 96.982368 way 34.200000 3 1 79.200000 way to 54 6 6
6 5 1 1 4 1 5 158 241 15 12 96.969261 to 45.000000 2 1 79.200000 way to 55 6 6
76.950000 signs, speedin' 15 3 5 1 1 1 1 3 104 82 42 18 93.283119 signs, 42.750000 6 1 76.950000 signs, speedin' 62 15 3
3 5 1 1 1 1 4 152 81 65 19 82.424248 speedin' 34.200000 8 1 76.950000 signs, speedin' 63 15 3
75.461538 Nobody's gonna 14 2 5 1 1 1 2 1 37 104 74 19 79.710175 Nobody's 39.461538 8 1 75.461538 Nobody's gonna 32 14 2
2 5 1 1 1 2 2 115 109 49 14 95.714142 gonna 36.000000 5 1 75.461538 Nobody's gonna 33 14 2
75.000000 No stop 7 5 5 1 1 1 1 1 37 82 23 14 95.939438 No 45.000000 2 1 75.000000fuzzcolumnsearch = df.text.ds_fuzz_multirow(fuzzsearch='Rocking BAND')
aa_npsum aa_weight aa_whole_text aa_whole_text_len aa_whole_text_len_difference
1 180.000000 rockin' band 12 0 rockin' 90.000000 7 1 180.000000 rockin' band 0 12 0
0 band 90.000000 4 1 180.000000 rockin' band 1 12 0
130.000000 Playin’ in 10 2 Playin’ 40.000000 7 1 130.000000 Playin’ in 12 10 2
2 in 90.000000 2 1 130.000000 Playin’ in 13 10 2
122.142857 promise land 12 0 promise 45.000000 7 1 122.142857 promise land 22 12 0
0 land 77.142857 4 1 122.142857 promise land 23 12 0
96.428571 Satan! Paid 11 1 Satan! 45.000000 6 1 96.428571 Satan! Paid 56 11 1
Regex search in multiple rows in a DataFrame - works with any DataFrame, not only tesseract DataFrames Only shows regex results that span over multiple rows!
level page_num block_num par_num line_num word_num left top width height conf text aa_regex_results aa_start aa_end
0 1 1 0 0 0 0 0 0 480 360 -1 <NA> <NA> <NA>
1 2 1 1 0 0 0 37 81 275 177 -1 <NA> <NA> <NA>
2 3 1 1 1 0 0 37 81 258 109 -1 <NA> <NA> <NA>
3 4 1 1 1 1 0 37 81 222 19 -1 <NA> <NA> <NA>
4 5 1 1 1 1 1 37 82 23 14 95.939438 No <NA> <NA> <NA>
5 5 1 1 1 1 2 66 84 33 16 96.663704 stop <NA> <NA> <NA>
6 5 1 1 1 1 3 104 82 42 18 93.283119 signs, <NA> <NA> <NA>
7 5 1 1 1 1 4 152 81 65 19 82.424248 speedin' <NA> <NA> <NA>
8 5 1 1 1 1 5 222 81 37 15 96.098083 limit <NA> <NA> <NA>
9 4 1 1 1 2 0 37 104 245 19 -1 <NA> <NA> <NA>
10 5 1 1 1 2 1 37 104 74 19 79.710175 Nobody's <NA> <NA> <NA>
11 5 1 1 1 2 2 115 109 49 14 95.714142 gonna <NA> <NA> <NA>
12 5 1 1 1 2 3 169 104 37 14 96.856789 slow <NA> <NA> <NA>
13 5 1 1 1 2 4 210 109 24 9 95.754181 me <NA> <NA> <NA>
14 5 1 1 1 2 5 239 104 43 14 96.090439 down <NA> <NA> <NA>
15 4 1 1 1 3 0 37 126 209 19 -1 <NA> <NA> <NA>
16 5 1 1 1 3 1 37 126 32 15 96.436874 Like Like a 64 70
17 5 1 1 1 3 2 74 131 8 10 96.436874 a Like a 64 70Regex search in multiple rows in a Series - works with any Series, not only tesseract Series Only shows regex results that span over multiple rows!
regexcolumnsearch = df.text.ds_regex_multirow( r'Like\s*\b\w+\b')
text aa_regex_results aa_start aa_end
0 <NA> <NA> <NA>
1 <NA> <NA> <NA>
2 <NA> <NA> <NA>
3 <NA> <NA> <NA>
4 No <NA> <NA> <NA>
5 stop <NA> <NA> <NA>
6 signs, <NA> <NA> <NA>
7 speedin' <NA> <NA> <NA>
8 limit <NA> <NA> <NA>
9 <NA> <NA> <NA>
10 Nobody's <NA> <NA> <NA>
11 gonna <NA> <NA> <NA>
12 slow <NA> <NA> <NA>
13 me <NA> <NA> <NA>
14 down <NA> <NA> <NA>
15 <NA> <NA> <NA>
16 Like Like a 64 70
17 a Like a 64 70
18 wheel, <NA> <NA> <NA>
19 gonna <NA> <NA> <NA>
df.text.ds_regex_multirow( r'mess.*dues')
23 Nobody's <NA> <NA> <NA>
24 gonna <NA> <NA> <NA>
25 mess mess me ‘round Hey Satan! Paid my dues 108 147
26 me mess me ‘round Hey Satan! Paid my dues 108 147
27 ‘round mess me ‘round Hey Satan! Paid my dues 108 147
28 mess me ‘round Hey Satan! Paid my dues 108 147
29 Hey mess me ‘round Hey Satan! Paid my dues 108 147
30 Satan! mess me ‘round Hey Satan! Paid my dues 108 147
31 Paid mess me ‘round Hey Satan! Paid my dues 108 147
32 my mess me ‘round Hey Satan! Paid my dues 108 147
33 dues mess me ‘round Hey Satan! Paid my dues 108 147
34 <NA> <NA> <NA>
35 <NA> <NA> <NA>
36 Playin’ <NA> <NA> <NA>