I wanted your help, I've been trying for a few months to make a code that finds a word in the image and returns the coordinates where that word is in the image.
I was trying this using OpenCV, OCR tesseract, but I was not successful, could someone here in the community help me?
I'll leave an image here as an example:
Here is something you can start with:
import pytesseract
from PIL import Image
pytesseract.pytesseract.tesseract_cmd = r'C:\<path-to-your-tesseract>\Tesseract-OCR\tesseract.exe'
img = Image.open("img.png")
data = pytesseract.image_to_data(img, output_type='dict')
boxes = len(data['level'])
for i in range(boxes):
if data['text'][i] != '':
print(data['left'][i], data['top'][i], data['width'][i], data['height'][i], data['text'][i])
If you have difficulties with installing pytesseract see: https://stackoverflow.com/a/53672281/18667225
Output:
153 107 277 50 Palavras
151 197 133 37 com
309 186 154 48 R/RR
154 303 126 47 Rato
726 302 158 47 Resto
154 377 144 50 Rodo
720 379 159 47 Arroz
152 457 160 48 Carro
726 457 151 46 Ferro
154 532 142 50 Rede
726 534 159 47 Barro
154 609 202 50 Parede
726 611 186 47 Barata
154 690 124 47 Faro
726 685 288 50 Beterraba
154 767 192 47 Escuro
726 766 151 47 Ferro
I managed to find the solution and I'll post it here for you:
import pytesseract
import cv2
from pytesseract import Output
pytesseract.pytesseract.tesseract_cmd = r'C:\<path-to-your-tesseract>\Tesseract-OCR\tesseract.exe'
filepath = 'image.jpg'
image = cv2.imread(filepath, 1)
# converting image to grayscale image
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# converting to binary image by Thresholding
# this step is necessary if you have a color image because if you skip this part
# then the tesseract will not be able to detect the text correctly and it will give an incorrect result
threshold_img = cv2.threshold(gray_image, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
# displays the image
cv2.imshow('threshold image', threshold_img)
# Holds the output window until the user presses a key
cv2.waitKey(0)
# Destroying windows present on the screen
cv2.destroyAllWindows()
# setting parameters for tesseract
custom_config = r'--oem 3 --psm 6'
# now feeding image to tesseract
details = pytesseract.image_to_data(threshold_img, output_type=Output.DICT, config=custom_config, lang='eng')
# Color
vermelho = (0, 0, 255)
#Exibe todas as chaves encontradas
print(details.keys())
print(details['text'])
# For in all found texts
for i in range(len(details['text'])):
# If it finds the text "UNIVERIDADE" it will print the coordinates, and draw a rectangle around the word
if details['text'][i] == 'UNIVERSIDADE':
print(details['text'][i])
print(f"left: {details['left'][i]}")
print(f"top: {details['top'][i]}")
print(f"width: {details['width'][i]}")
print(f"height: {details['height'][i]}")
cv2.rectangle(image, (details['left'][i], details['top'][i]), (details['left'][i]+details['width'][i], details['top'][i]+details['height'][i]), vermelho)
I have a sqlite database file with several columns. One of the columns has a JSON dictionary (with two keys) embedded in it. I want to extract the JSON column to a data frame in R that shows each key in a separate column.
I tried rjson::fromJSON, but it reads only the first item. Is there a trick that I'm missing?
Here's an example that mimics my problem:
> eg <- as.vector(c("{\"3x\": 20, \"6y\": 23}", "{\"3x\": 60, \"6y\": 50}"))
> fromJSON(eg)
$3x
[1] 20
$6y
[1] 23
The desired output is something like:
# a data frame for both variables
3x 6y
1 20 23
2 60 50
or,
# a data frame for each variable
3x
1 20
2 60
6y
1 23
2 50
What you are looking for is actually a combination of lapply and some application of rbind or related.
I'll extend your data a little, just to have more than 2 elements.
eg <- c("{\"3x\": 20, \"6y\": 23}",
"{\"3x\": 60, \"6y\": 50}",
"{\"3x\": 99, \"6y\": 72}")
library(jsonlite)
Using base R, we can do
do.call(rbind.data.frame, lapply(eg, fromJSON))
# X3x X6y
# 1 20 23
# 2 60 50
# 3 99 72
You might be tempted to do something like Reduce(rbind, lapply(eg, fromJSON)), but the notable difference is that in the Reduce model, rbind is called "N-1" times, where "N" is the number of elements in eg; this results in a LOT of copying of data, and though it might work alright with small "N", it scales horribly. With the do.call option, rbind is called exactly once.
Notice that the column labels have been R-ized, since data.frame column names should not start with numbers. (It is possible, but generally discouraged.)
If you're confident that all substrings will have exactly the same elements, then you may be good here. If there's a chance that there will be a difference at some point, perhaps
eg <- c(eg, "{\"3x\": 99}")
then you'll notice that the base R solution no longer works by default.
do.call(rbind.data.frame, lapply(eg, fromJSON))
# Error in (function (..., deparse.level = 1, make.row.names = TRUE, stringsAsFactors = default.stringsAsFactors()) :
# numbers of columns of arguments do not match
There may be techniques to try to normalize the elements such that you can be assured of matches. However, if you're not averse to a tidyverse package:
library(dplyr)
eg2 <- bind_rows(lapply(eg, fromJSON))
eg2
# # A tibble: 4 × 2
# `3x` `6y`
# <int> <int>
# 1 20 23
# 2 60 50
# 3 99 72
# 4 99 NA
though you cannot call it as directly with the dollar-method, you can still use [[ or backticks.
eg2$3x
# Error: unexpected numeric constant in "eg2$3"
eg2[["3x"]]
# [1] 20 60 99 99
eg2$`3x`
# [1] 20 60 99 99
From ELF Documentation:
SHT_STRTAB
The section holds a string table. An object file may have multiple string table sections. See ‘‘String Table’’ below for details.
(Note: I didn't notice any information regarding multiple string table sections in ‘‘String Table’’ paragraph)
Does multiple string table sections mean string table for the section headers and string table for the object file itself?
There is no mention in the documentation regarding how to read a string if there are multiple string table for the object itself (.strtab).
Any clarification about the subject appreciated.
The manpage only has a overview/summary of (some parts of) the ELF file format, you might want to look at the System V ABI spec.
Explanation
Linking view
An ELF file has multiple string tables. Usually you have 3-4 string tables:
One string table (usually called .shstrtab) is used for section names. All section names (in the section header table) are taken from a single string table. This string table is identified by its index in the section header table: the index of the section name string table is indicated in the ELF header (e_shstrndx).
Another string table (usually called .strtab) is used for the full symbol table (.symtab). The same string table is used by the .dynamic section.
Another string table (usually called .dynstr) is used for the minirmal symbol table (.dynsym).
Another string table is used for
For a given symbol table section, the section used as string table is indicated in the sh_link field of the section header table (see Fig 4-12 of the system V ABI spec).
Execution view
For the execution view (the program header table), the address of the string table used for the symbol table (DT_SYMTAB) is given in the DT_STRTAB entry of the dynamic section.
Example
Linking view
This is a hello world program (shown with readelf -a).
.shtrtab
The ELF header:
ELF Header:
Magic: 7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00
Class: ELF64
Data: 2's complement, little endian
Version: 1 (current)
OS/ABI: UNIX - System V
ABI Version: 0
Type: EXEC (Executable file)
Machine: Advanced Micro Devices X86-64
Version: 0x1
Entry point address: 0x4003c0
Start of program headers: 64 (bytes into file)
Start of section headers: 4624 (bytes into file)
Flags: 0x0
Size of this header: 64 (bytes)
Size of program headers: 56 (bytes)
Number of program headers: 8
Size of section headers: 64 (bytes)
Number of section headers: 30
Section header string table index: 27
tells us that the section names are in section 27. Conveniently enough this is .shtrtab:
Section Headers:
[Nr] Name Type Address Offset
Size EntSize Flags Link Info Align
[...]
[27] .shstrtab STRTAB 0000000000000000 000008e0
0000000000000108 0000000000000000 0 0 1
.dynstr
For .dynsym we have:
[ 5] .dynsym DYNSYM 0000000000400280 00000280
0000000000000048 0000000000000018 A 6 1 8
^
HERE
Its names are taken from section 6 which is .dynstr:
[ 6] .dynstr STRTAB 00000000004002c8 000002c8
0000000000000038 0000000000000000 A 0 0 1
This string table is used by other sections as well:
[ 8] .gnu.version_r VERNEED 0000000000400308 00000308
0000000000000020 0000000000000000 A 6 1 8
[21] .dynamic DYNAMIC 0000000000600698 00000698
00000000000001d0 0000000000000010 WA 6 0 8
.strtab
For .symtab:
[28] .symtab SYMTAB 0000000000000000 000009e8
0000000000000600 0000000000000018 29 45 8
^
HERE
the names are taken from section 29 which happens to be .strtab:
[29] .strtab STRTAB 0000000000000000 00000fe8
0000000000000224 0000000000000000 0 0 1
Execution view
Dynamic section at offset 0x698 contains 24 entries:
Tag Type Name/Value
0x0000000000000001 (NEEDED) Shared library: [libc.so.6]
0x000000000000000c (INIT) 0x400370
0x000000000000000d (FINI) 0x400544
0x0000000000000019 (INIT_ARRAY) 0x600680
0x000000000000001b (INIT_ARRAYSZ) 8 (bytes)
0x000000000000001a (FINI_ARRAY) 0x600688
0x000000000000001c (FINI_ARRAYSZ) 8 (bytes)
0x000000006ffffef5 (GNU_HASH) 0x400260
0x0000000000000005 (STRTAB) 0x4002c8 <= HERE
0x0000000000000006 (SYMTAB) 0x400280
0x000000000000000a (STRSZ) 56 (bytes)
0x000000000000000b (SYMENT) 24 (bytes)
0x0000000000000015 (DEBUG) 0x0
0x0000000000000003 (PLTGOT) 0x600870
0x0000000000000002 (PLTRELSZ) 48 (bytes)
0x0000000000000014 (PLTREL) RELA
0x0000000000000017 (JMPREL) 0x400340
0x0000000000000007 (RELA) 0x400328
0x0000000000000008 (RELASZ) 24 (bytes)
0x0000000000000009 (RELAENT) 24 (bytes)
0x000000006ffffffe (VERNEED) 0x400308
0x000000006fffffff (VERNEEDNUM) 1
0x000000006ffffff0 (VERSYM) 0x400300
0x0000000000000000 (NULL) 0x0
The string table for dynamic linking is located at 0x4002c8 in the program memory.
Note: this is .dynstr.
I have a code written in C++ that outputs a .csv file with data in three columns (Time, Force, Height). I want to plot the data using Octave, or else use the octave function plot in the C++ file (I'm aware this is possible but I don't necessarily need to do it this way).
Right now I have the simple .m file:
filename = linear_wave_loading.csv;
M = csvread(filename);
Just to practice bringing this file into octave (will try and plot after)
I am getting this error.
error: dlmread: error parsing range
What is the correct method to load .csv files into octave?
Edit: Here is the first few lines of my .csv file
Wavelength= 88.7927 m
Time Height Force(KN/m)
0 -20 70668.2
0 -19 65875
0 -18 61411.9
0 -17 57256.4
Thanks in advance for your help.
Using octave 3.8.2
>> format long g
>> dlmread ('test.csv',' ',2,0)
ans =
0 0 0 -20 70668.2
0 0 0 -19 65875
0 0 0 -18 61411.9
0 0 0 -17 57256.4
General, use dlmread if your value separator is not a comma. Furthermore, you have to skip the two headlines.
Theoretical dlmread works with tab separated values too '\t', but this failes with you given example, because of the discontinuous tab size (maybe it's just a copy paste problem), so taking one space ' ' as separator is a workaround.
you should better save your .csv file comma separated
Wavelength= 88.7927 m
Time Height Force(KN/m)
0, -20, 70668.2
0, -19, 65875
0, -18, 61411.9
0, -17, 57256.4
Then you can easily do dlmread('file.csv',',',2,0).
You can try my csv2cell(not to be confused with csv2cell from io package!) function (never tried it < 3.8.0).
>> str2double(reshape(csv2cell('test.csv', ' +',2),3,4))'
ans =
0 -20 70668.2
0 -19 65875
0 -18 61411.9
0 -17 57256.4
Usually it reshaped successful automatically, but in case of space seperators, it often failed, so you have to reshape it by your self (and convert to double in any case).
And when you need your headline
>> reshape(csv2cell('test.csv', ' +',1),3,5)'
ans =
{
[1,1] = Time
[2,1] = +0
[3,1] = +0
[4,1] = +0
[5,1] = +0
[1,2] = Height
[2,2] = -20
[3,2] = -19
[4,2] = -18
[5,2] = -17
[1,3] = Force(KN/m)
[2,3] = 70668.2
[3,3] = 65875
[4,3] = 61411.9
[5,3] = 57256.4
}
But take care, then everything is a string in your cell.
Your not storing you .csv filename as a string.
Try:
filename = 'linear_wave_loading.csv';
Given the output of git ... --stat:
3 files changed, 72 insertions(+), 21 deletions(-)
3 files changed, 27 insertions(+), 4 deletions(-)
4 files changed, 164 insertions(+), 0 deletions(-)
9 files changed, 395 insertions(+), 0 deletions(-)
1 files changed, 3 insertions(+), 2 deletions(-)
1 files changed, 1 insertions(+), 1 deletions(-)
2 files changed, 57 insertions(+), 0 deletions(-)
10 files changed, 189 insertions(+), 230 deletions(-)
3 files changed, 111 insertions(+), 0 deletions(-)
8 files changed, 61 insertions(+), 80 deletions(-)
I wanted to produce the sum of the numeric columns but preserve the formatting of the line. In the interest of generality, I produced this awk script that automatically sums any numeric columns and produces a summary line:
{
for (i = 1; i <= NF; ++i) {
if ($i + 0 != 0) {
numeric[i] = 1;
total[i] += $i;
}
}
}
END {
# re-use non-numeric columns of last line
for (i = 1; i <= NF; ++i) {
if (numeric[i])
$i = total[i]
}
print
}
Yielding:
44 files changed, 1080 insertions(+), 338 deletions(-)
Awk has several features that simplify the problem, like automatic string->number conversion, all arrays as associative arrays, and the ability to overwrite auto-split positional parameters and then print the equivalent lines.
Is there a better language for this hack?
Perl - 47 char
Inspired by ChristopheD's awk solution. Used with the -an command-line switch. 43 chars + 4 chars for the command-line switch:
$i-=#a=map{($b[$i++]+=$_)||$_}#F}{print"#a"
I can get it to 45 (41 + -ap switch) with a little bit of cheating:
$i=0;$_="Ctrl-M#{[map{($b[$i++]+=$_)||$_}#F]}"
Older, hash-based 66 char solution:
#a=(),s#(\d+)(\D+)#$b{$a[#a]=$2}+=$1#gefor<>;print map$b{$_}.$_,#a
Ruby — 87
puts ' '+[*$<].map(&:split).inject{|i,j|[0,3,5].map{|k|i[k]=i[k].to_i+j[k].to_i};i}*' '
Python - 101 chars
import sys
print" ".join(`sum(map(int,x))`if"A">x[0]else x[0]for x in zip(*map(str.split,sys.stdin)))'
Using reduce is longer at 126 chars
import sys
print" ".join(reduce(lambda X,Y:[str(int(x)+int(y))if"A">x[0]else x for x,y in zip(X,Y)],map(str.split,sys.stdin)))
AWK - 63 characters
(in a bash script, $1 is the filename provided as command line argument):
awk -F' ' '{x+=$1;y+=$4;z+=$6}END{print x,$2,$3,y,$5,z,$7}' $1
One could of course also pipe the input in (would save another 3 characters when allowed).
This problem is not challenging or difficult... it is "cute" though.
Here is solution in Python:
import sys
r = []
for s in sys.stdin:
r = map(lambda x,y:(x or 0)+int(y) if y.isdigit() else y, r, s.split())
print ' '.join(map(str, r))
What does it do... it keeps tally in r while proceeding line by line. Splits the line, then for each element of the list, if it is a number, adds it to the tally or keeps it as string. At the end they all get re-mapped to string and merged with spaces in between to be printed.
Alternative, more "algebraic" implementation, if we did not care about reading all input at once:
import sys
def totalize(l):
try: r = str(sum(map(int,l)))
except: r = l[-1]
return r
print ' '.join(map(totalize, zip(*map(str.split, sys.stdin))))
What does this one do? totalize() takes a list of strings and tries to calculate sum of the numbers; if that fails, it simply returns the last one. zip() is fed with a matrix that is list of rows, each of them being list of column items in the row - zip transposes the matrix so it turns into list of column items and then totalize is invoked on each column and the results are joined as before.
At the expense of making your code slightly longer, I moved the main parsing into the BEGIN clause so the main clause is only processing numeric fields. For a slightly larger input file, I was able to measure a significant improvement in speed.
BEGIN {
getline
for (i = 1; i <= NF; ++i) {
# need to test for 0, too, in this version
if ($i == 0 || $i + 0 != 0) {
numeric[i] = 1;
total[i] = $i;
}
}
}
{
for (i in numeric) total[i] += $i
}
END {
# re-use non-numeric columns of last line
for (i = 1; i <= NF; ++i) {
if (numeric[i])
$i = total[i]
}
print
}
I made a test file using your data and doing paste file file file ... and cat file file file ... so that the result had 147 fields and 1960 records. My version took about 1/4 as long as yours. On the original data, the difference was not measurable.
JavaScript (Rhino) - 183 154 139 bytes
Golfed:
x=[n=0,0,0];s=[];readFile('/dev/stdin').replace(/(\d+)(\D+)/g,function(a,b,c){x[n]+=+b;s[n++]=c;n%=3});print(x[0]+s[0]+x[1]+s[1]+x[2]+s[2])
Readable-ish:
x=[n=0,0,0];
s=[];
readFile('/dev/stdin').replace(/(\d+)(\D+)/g,function(a,b,c){
x[n]+=+b;
s[n++]=c;
n%=3
});
print(x[0]+s[0]+x[1]+s[1]+x[2]+s[2]);
PHP 152 130 Chars
Input:
$i = "
3 files changed, 72 insertions(+), 21 deletions(-)
3 files changed, 27 insertions(+), 4 deletions(-)
4 files changed, 164 insertions(+), 0 deletions(-)
9 files changed, 395 insertions(+), 0 deletions(-)
1 files changed, 3 insertions(+), 2 deletions(-)
1 files changed, 1 insertions(+), 1 deletions(-)
2 files changed, 57 insertions(+), 0 deletions(-)
10 files changed, 189 insertions(+), 230 deletions(-)
3 files changed, 111 insertions(+), 0 deletions(-)
8 files changed, 61 insertions(+), 80 deletions(-)";
Code:
$a = explode(" ", $i);
foreach($a as $k => $v){
if($k % 7 == 0)
$x += $v;
if(3-$k % 7 == 0)
$y += $v;
if(5-$k % 7 == 0)
$z += $v;
}
echo "$x $a[1] $a[2] $y $a[4] $z $a[6]";
Output:
44 files changed, 1080 insertions(+), 338 deletions(-)
Note: explode() will require that there is a space char before the new line.
Haskell - 151 135 bytes
import Char
c a b|all isDigit(a++b)=show$read a+read b|True=a
main=interact$unwords.foldl1(zipWith c).map words.filter(not.null).lines
... but I'm sure it can be done better/smaller.
Lua, 140 bytes
I know Lua isn't the best golfing language, but compared by the size of the runtimes, it does pretty well I think.
f,i,d,s=0,0,0,io.read"*a"for g,a,j,b,e,c in s:gmatch("(%d+)(.-)(%d+)(.-)(%d+)(.-)")do f,i,d=f+g,i+j,d+e end print(table.concat{f,a,i,b,d,c})
PHP, 176 166 164 159 158 153
for($a=-1;$a<count($l=explode("
",$i));$r=explode(" ",$l[++$a]))for($b=-1;$b<count($r);$c[++$b]=is_numeric($r[$b])?$c[$b]+$r[$b]:$r[$b]);echo join(" ",$c);
This would, however, require the whole input in $i... A variant with $i replaced with $_POST["i"] so it would be sent in a textarea... Has 162 chars:
for($a=-1;$a<count($l=explode("
",$_POST["i"]));$r=explode(" ",$l[$a++]))for($b=0;$b<count($r);$c[$b]=is_numeric($r[$b])?$c[$b]+$r[$b]:$r[$b])$b++;echo join(" ",$c);
This is a version with
NO HARDCODED COLUMNS