From d7d3502a256bf132a3d2f07d6fdda387b07b8efa Mon Sep 17 00:00:00 2001
From: Johannes Zillmann <jzillmann@googlemail.com>
Date: Sun, 28 Mar 2021 10:21:26 +0200
Subject: [PATCH] Fix processing pdfs with no page numbers

---
 core/src/transformer/RemoveRepetitiveItems.ts | 10 +++---
 core/test/Files.test.ts                       |  5 ++-
 examples/README.md                            |  3 ++
 .../removeRepetitiveItems.json                | 32 +++++++++++++++++++
 4 files changed, 45 insertions(+), 5 deletions(-)
 create mode 100644 examples/compressed.tracemonkey-pldi-09/removeRepetitiveItems.json

diff --git a/core/src/transformer/RemoveRepetitiveItems.ts b/core/src/transformer/RemoveRepetitiveItems.ts
index 4fb0325..6b62dfa 100644
--- a/core/src/transformer/RemoveRepetitiveItems.ts
+++ b/core/src/transformer/RemoveRepetitiveItems.ts
@@ -161,10 +161,12 @@ function filterOutIncompatibleVariant(options: PageNumber[], nextPageLines: Page
   let remainingOptions = [...options];
   while (remainingOptions.length > 1 && index < nextPageLines.length) {
     const nextPageNumbers = possiblePageNumbers(nextPageLines[index]);
-    remainingOptions = remainingOptions.filter((option) => {
-      const maxDistance = nextPageNumbers[0].pageIndex - option.pageIndex;
-      return nextPageNumbers.find((nextPageNum) => nextPageNum.pageNumber - option.pageNumber <= maxDistance);
-    });
+    if (nextPageNumbers.length > 0) {
+      remainingOptions = remainingOptions.filter((option) => {
+        const maxDistance = nextPageNumbers[0].pageIndex - option.pageIndex;
+        return nextPageNumbers.find((nextPageNum) => nextPageNum.pageNumber - option.pageNumber <= maxDistance);
+      });
+    }
     index++;
   }
   return remainingOptions;
diff --git a/core/test/Files.test.ts b/core/test/Files.test.ts
index 1335427..4f556b6 100644
--- a/core/test/Files.test.ts
+++ b/core/test/Files.test.ts
@@ -19,7 +19,10 @@ const pipeline = new PdfPipeline(parser, transformers);
 
 const folder = '../examples';
 const files = fs.readdirSync(folder).filter((file) => file.endsWith('.pdf'));
-const urls = ['https://homepages.cwi.nl/~lex/files/dict.pdf'];
+const urls = [
+  'https://homepages.cwi.nl/~lex/files/dict.pdf',
+  'https://github.com/mozilla/pdf.js/raw/master/web/compressed.tracemonkey-pldi-09.pdf',
+];
 const downloadCache = 'node_modules/.cache/files';
 
 expect.extend({ toMatchFile });
diff --git a/examples/README.md b/examples/README.md
index 1f9c5f1..1650b09 100644
--- a/examples/README.md
+++ b/examples/README.md
@@ -36,6 +36,9 @@ _(PDFs which entered `public domain` or have a otherwise permissive license like
 ## PDFs not stored but paritally tested
 
 - https://homepages.cwi.nl/~lex/files/dict.pdf
+  - Page numbers with current chapter
+- https://github.com/mozilla/pdf.js/raw/master/web/compressed.tracemonkey-pldi-09.pdf
+  - No page numbers
 
 # Known transformatino problems
 
diff --git a/examples/compressed.tracemonkey-pldi-09/removeRepetitiveItems.json b/examples/compressed.tracemonkey-pldi-09/removeRepetitiveItems.json
new file mode 100644
index 0000000..3a27574
--- /dev/null
+++ b/examples/compressed.tracemonkey-pldi-09/removeRepetitiveItems.json
@@ -0,0 +1,32 @@
+{
+  "pages": 14,
+  "items": 2430,
+  "groupedItems": 1497,
+  "changes": 0,
+  "schema": [
+    {
+      "name": "line"
+    },
+    {
+      "name": "x"
+    },
+    {
+      "name": "y"
+    },
+    {
+      "name": "str"
+    },
+    {
+      "name": "fontName"
+    },
+    {
+      "name": "dir"
+    },
+    {
+      "name": "width"
+    },
+    {
+      "name": "height"
+    }
+  ]
+}
\ No newline at end of file