Extracts data from GSF PDFs for SMART facilitators. Data is extracted using Tabula (with its python bindings), then cleaned up. Since PDF data extraction is not perfect, the script collections emails (strings containing '@') and outputs the set of emails for which no entry was extracted; these will have to be extracted by hand.
- python (tested in 3.6)
- java (7 or 8)
- tabula-py
- Download and install python and Java, if you don't have them
- (optional) Install virtualenv and configure an environment
pip install virtualenvvirtualenv pdfpdf\Scripts\activate(usedeactivateto exit the virtualenv)
- Install tabula-py:
pip install tabula-pyNOTE: tabula-py relies on pandas; it is strongly recommended that you install this within a virtualenv
python smart_extractor.py input.pdf
This will extract data from input.pdf, saving the extracted data in output.csv and the missed emails in missed.csv.
If output.csv or missed.csv exist, you can supply different output names with -o and -m, or use -f to overwrite
existing files.
python smart_extractor.py input1.pdf path/to/input2.pdf directory/of/pdfs/
You can supply multiple PDF files, as well as directories containing PDFs. The results are collected and written to output.csv and missed.csv.
usage: smart_extractor.py [-h] [-o OUTPUT] [-m MISSED] [-f] [-q]
input [input ...]
Extract GSF data from PDF. Good entries are saved in output.csv; entries that
couldn't be procssed are saved in output-missed.csv.
positional arguments:
input a PDF file, or directory of PDF files, to process
optional arguments:
-h, --help show this help message and exit
-o OUTPUT, --output OUTPUT
filename for extracted entries
-m MISSED, --missed MISSED
filename for emailed not found in extracted entries
-f, --force force overwrite of existing output files
-q, --quiet don't print processing info to the command line