seomoz / dragnet_data Goto Github PK

I was trying to use the data for some experiments, but when reading it directly with open in python3, in encountered an encoding error for file R121.html:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xcc in position 1213: invalid continuation byte

I wrote a small script to check the encoding of the failing files using the chardet utility. The code is below.

for f in *.html; do
        encoding=$(file -i $f | cut -d"=" -f 2)  # get the mime encoding
        if [ "$encoding" != "us-ascii" ] && [ "$encoding" != "utf-8" ]; then
                res=$(chardetect $f)  # try to detect it otherwise
                encoding=$(echo $res | cut -d" " -f 2)
                echo $res
        fi
done

And produces the following result:

R121.html: Windows-1254 with confidence 0.5434633906826465
R17.html: ISO-8859-1 with confidence 0.73
R736.html: Windows-1252 with confidence 0.73
R757.html: Windows-1252 with confidence 0.73
R826.html: Windows-1252 with confidence 0.73
R827.html: Windows-1252 with confidence 0.73
T156.html: windows-1251 with confidence 0.7538428528079772
T19.html: Windows-1252 with confidence 0.73
T2.html: Windows-1254 with confidence 0.5434729438118417
T31.html: Windows-1254 with confidence 0.5239184224706976
T97.html: ISO-8859-1 with confidence 0.73

These inconsistencies are not major and I managed to fix them afterwards with a few changes to the detection script, but a few failed even with recode(R121.html, T19.html, T2.html, T31.html) and I had to remove them. Here is the script I used to convert the inconsistent ones.

for f in *.html; do
        encoding=$(file -i $f | cut -d"=" -f 2)  # get the mime encoding
        if [ "$encoding" != "us-ascii" ] && [ "$encoding" != "utf-8" ]; then
                res=$(chardetect $f)  # try to detect it otherwise
                encoding=$(echo $res | cut -d" " -f 2)
                echo $res - CONVERTING TO UTF-8
                recode ${encoding}..utf-8 $f
        fi
done

This might be an issue on my part, maybe, so I'm curious if this is something that came to your attention before.

Hi, I'm currently compiling additional, more modern html documents with gold standard content + comments for use in training dragnet models, and I have a few questions:

• Should I consider the text of embedded tweets, posts, quotations, image captions, and other rich media to be content?
• Should I include author byline, pubdate, etc. at the start of an article as content? What about typical addenda at the bottom of the article?
• Should I include non-English language content?

Thanks for your help!