crude djvu XML parsing

1 files changed, 16 insertions, 0 deletions

diff --git a/tests/test_djvu_parse.py b/tests/test_djvu_parse.py
new file mode 100644
index 0000000..66c2804
--- /dev/null
+++ b/tests/test_djvu_parse.py
@@ -0,0 +1,16 @@
+
+from io import StringIO
+from fatcat_scholar.djvu import djvu_extract_leaf_texts
+
+
+def test_djvu_extract_leaf_texts():
+
+    # https://archive.org/details/ERIC_ED441501
+    with open('tests/files/ERIC_ED441501_djvu.xml', 'r') as f:
+        blob = f.read()
+
+    leaves = djvu_extract_leaf_texts(StringIO(blob), [3,6])
+    assert 3 in leaves
+    assert 6 in leaves
+    assert "2. Original cataloging tools" in leaves[3]
+    assert len(leaves) == 2