lixiangnlp/nytimes-corpus-extractor

Extract all the fields from the NY Times Corpus to a csv

Extract All the Fields from the New York Times Corpus to a CSV

The New York Times Corpus is a collection of 1.8 million articles published between 1987 and 2007 along with a fair bit of meta data. For more details about The NY Times Corpus, see https://catalog.ldc.upenn.edu/LDC2008T19.

Once you have the NY Times Corpus, unzip it to a folder. And then run the script. Script produces a csv and text files containing story text.

Requirements

lxml 3.1.1

Usage

Usage: nytextract.py [options] <xml directory>

Options:
  -h, --help            show this help message and exit
  -a, --append          Append if existing (default: False)
  -o OUTFILE, --out=OUTFILE
                        CSV output file (default: outfile.csv)
  -d OUTDIR, --dir=OUTDIR
                        Text output directory (default: text)

Example To process all XML files in the folder 2000 (carrying files from year 2000):

python nytextract.py -o 2000.csv -d text 2000

The script will generate a CSV "2000.csv". Story text files will be stored in a folder "text." This folder will have the exact same structure as the folder '2000.'

License

Scripts are released under the MIT License.