From 94912e739c51d2fa4d5f9de878d0b0f0544a4459 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Wed, 12 Feb 2020 19:40:55 -0800
Subject: pdftrio basic python code

This is basically just a copy/paste of GROBID code, only simpler!
---
 python/pdftrio_tool.py | 118 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 118 insertions(+)
 create mode 100755 python/pdftrio_tool.py

(limited to 'python/pdftrio_tool.py')

diff --git a/python/pdftrio_tool.py b/python/pdftrio_tool.py
new file mode 100755
index 0000000..843c214
--- /dev/null
+++ b/python/pdftrio_tool.py
@@ -0,0 +1,118 @@
+#!/usr/bin/env python3
+
+"""
+Basically just a copy of grobid_tool.py, but for PDF classification instead of
+text extraction.
+
+Example of large parallel run, locally:
+
+    cat /srv/sandcrawler/tasks/something.cdx | pv -l | parallel -j30 --pipe ./pdftrio_tool.py --kafka-env prod --kafka-hosts wbgrp-svc263.us.archive.org:9092,wbgrp-svc284.us.archive.org:9092,wbgrp-svc285.us.archive.org:9092 --kafka-mode --pdftrio-host http://localhost:3939 -j0 classify-pdf-json -
+"""
+
+import sys
+import json
+import argparse
+import datetime
+
+from sandcrawler import *
+
+
+def run_classify_pdf_json(args):
+    pdftrio_client = PdfTrioClient(host_url=args.pdftrio_host)
+    wayback_client = WaybackClient()
+    if args.jobs > 1:
+        worker = PdfTrioWorker(pdftrio_client, wayback_client, sink=None)
+        multi_worker = MultiprocessWrapper(worker, args.sink)
+        pusher = JsonLinePusher(multi_worker, args.json_file, batch_size=args.jobs)
+    else:
+        worker = PdfTrioWorker(pdftrio_client, wayback_client, sink=args.sink)
+        pusher = JsonLinePusher(worker, args.json_file)
+    pusher.run()
+
+def run_classify_pdf_cdx(args):
+    pdftrio_client = PdfTrioClient(host_url=args.pdftrio_host)
+    wayback_client = WaybackClient()
+    if args.jobs > 1:
+        worker = PdfTrioWorker(pdftrio_client, wayback_client, sink=None)
+        multi_worker = MultiprocessWrapper(worker, args.sink)
+        pusher = CdxLinePusher(
+            multi_worker,
+            args.cdx_file,
+            filter_http_statuses=[200, 226],
+            filter_mimetypes=['application/pdf'],
+            batch_size=args.jobs,
+        )
+    else:
+        worker = PdfTrioWorker(pdftrio_client, wayback_client, sink=args.sink)
+        pusher = CdxLinePusher(
+            worker,
+            args.cdx_file,
+            filter_http_statuses=[200, 226],
+            filter_mimetypes=['application/pdf'],
+        )
+    pusher.run()
+
+def run_classify_pdf_zipfile(args):
+    pdftrio_client = PdfTrioClient(host_url=args.pdftrio_host)
+    worker = PdfTrioBlobWorker(pdftrio_client, sink=args.sink)
+    pusher = ZipfilePusher(worker, args.zip_file)
+    pusher.run()
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+    parser.add_argument('--kafka-mode',
+        action='store_true',
+        help="send output to Kafka (not stdout)")
+    parser.add_argument('--kafka-hosts',
+        default="localhost:9092",
+        help="list of Kafka brokers (host/port) to use")
+    parser.add_argument('--kafka-env',
+        default="dev",
+        help="Kafka topic namespace to use (eg, prod, qa, dev)")
+    parser.add_argument('-j', '--jobs',
+        default=8, type=int,
+        help="parallelism for batch CPU jobs")
+    parser.add_argument('--pdftrio-host',
+        default="http://pdftrio.qa.fatcat.wiki",
+        help="pdftrio API host/port")
+    subparsers = parser.add_subparsers()
+
+    sub_classify_pdf_json = subparsers.add_parser('classify-pdf-json',
+        help="for each JSON line with CDX info, fetches PDF and does pdftrio classify_pdfion")
+    sub_classify_pdf_json.set_defaults(func=run_classify_pdf_json)
+    sub_classify_pdf_json.add_argument('json_file',
+        help="JSON file to import from (or '-' for stdin)",
+        type=argparse.FileType('r'))
+
+    sub_classify_pdf_cdx = subparsers.add_parser('classify-pdf-cdx',
+        help="for each CDX line, fetches PDF and does pdftrio classify_pdfion")
+    sub_classify_pdf_cdx.set_defaults(func=run_classify_pdf_cdx)
+    sub_classify_pdf_cdx.add_argument('cdx_file',
+        help="CDX file to import from (or '-' for stdin)",
+        type=argparse.FileType('r'))
+
+    sub_classify_pdf_zipfile = subparsers.add_parser('classify-pdf-zipfile',
+        help="opens zipfile, iterates over PDF files inside and does pdftrio classify_pdf for each")
+    sub_classify_pdf_zipfile.set_defaults(func=run_classify_pdf_zipfile)
+    sub_classify_pdf_zipfile.add_argument('zip_file',
+        help="zipfile with PDFs to classify",
+        type=str)
+
+    args = parser.parse_args()
+    if not args.__dict__.get("func"):
+        print("tell me what to do!")
+        sys.exit(-1)
+
+    args.sink = None
+    if args.kafka_mode:
+        produce_topic = "sandcrawler-{}.pdftrio-output".format(args.kafka_env)
+        print("Running in kafka output mode, publishing to {}\n".format(produce_topic))
+        args.sink = KafkaSink(kafka_hosts=args.kafka_hosts,
+            produce_topic=produce_topic)
+
+    args.func(args)
+
+if __name__ == '__main__':
+    main()
-- 
cgit v1.2.3