From 2339dd531becb3e7578d5fd0eacbb735ab2a414c Mon Sep 17 00:00:00 2001
From: zPlus <zplus@peers.community>
Date: Mon, 6 Jan 2025 14:55:02 +0100
Subject: [PATCH] pdf2txt read/write from/to pipe.

Edit the pdf2txt script to read the plaintext from stdin, and write
jsonld to stdout.
---
 scripts/pdf2txt/README     |  6 +++---
 scripts/pdf2txt/pdf2txt.py | 28 ++++++++++++++++++++++++++++
 scripts/pdf2txt/txt2rdf.py | 27 ---------------------------
 3 files changed, 31 insertions(+), 30 deletions(-)
 create mode 100755 scripts/pdf2txt/pdf2txt.py
 delete mode 100755 scripts/pdf2txt/txt2rdf.py

diff --git a/scripts/pdf2txt/README b/scripts/pdf2txt/README
index b7a0a57..ee19bd7 100644
--- a/scripts/pdf2txt/README
+++ b/scripts/pdf2txt/README
@@ -1,6 +1,6 @@
 Use pdftotext to convert library PDF files into plaintext.
 https://dokk.org/manpages/debian/12/poppler-utils/pdftotext.1.en
 
-    mkdir nodes txt
-    for file in *.pdf; do pdftotext -layout "$file" - > txt/"$(basename "$file").txt"; done
-    for file in txt/*; do ./txt2rdf.py "$(basename "$file")"; done
+    mkdir nodes
+    export PDF_FOLDER=
+    i=0; for file in $(find $PDF_FOLDER -type f -name *.pdf); do i=$(( i + 1 )); pdftotext -layout "$file" - | ./pdf2rdf.py $file > nodes/$i.jsonld; done
diff --git a/scripts/pdf2txt/pdf2txt.py b/scripts/pdf2txt/pdf2txt.py
new file mode 100755
index 0000000..56d6865
--- /dev/null
+++ b/scripts/pdf2txt/pdf2txt.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+
+# Take the plaintext as input and output RDF
+
+import json
+import os
+import sys
+from pathlib import Path
+
+assert 'PDF_FOLDER' in os.environ
+assert len(sys.argv) == 2
+PDF_FOLDER = os.environ['PDF_FOLDER']
+filename = sys.argv[1]
+assert filename.endswith('.pdf')
+plaintext = sys.stdin.read()
+
+file_id = Path(filename).relative_to(Path(PDF_FOLDER))
+
+node = {
+    '@context': {
+        'blob': 'dokk:vocab:blob:'
+    },
+
+    '@id': f'file:/pdf/{file_id}',
+    'blob:pdftotext': plaintext
+}
+
+print(json.dumps(node, indent=4, ensure_ascii=False))
diff --git a/scripts/pdf2txt/txt2rdf.py b/scripts/pdf2txt/txt2rdf.py
deleted file mode 100755
index 941f8ba..0000000
--- a/scripts/pdf2txt/txt2rdf.py
+++ /dev/null
@@ -1,27 +0,0 @@
-#!/usr/bin/env python3
-
-# Take the plaintext as input and output RDF
-
-import json
-import sys
-
-filename = sys.argv[1]
-
-# Remove file extension
-assert filename.endswith('.pdf.txt')
-id = filename[:-8]
-
-with open(f'txt/{filename}', 'r') as file:
-    plaintext = file.read()
-
-node = {
-    '@context': {
-        'blob': 'dokk:vocab:blob:'
-    },
-
-    '@id': f'file:/pdf/{id}.pdf',
-    'blob:pdftotext': plaintext
-}
-
-with open(f'nodes/{id}.jsonld', 'w') as file:
-    json.dump(node, file, indent=4, ensure_ascii=False)