Implement OneByteIdentityH/V encoding cmap

Original: euske/pdfminer#179
eladkehat · May 1, 2019 · a24b226 · a24b226
1 parent a1421ba
commit a24b226
Show file tree

Hide file tree

Showing 7 changed files with 48 additions and 2 deletions.
diff --git a/pdfminer/cmapdb.py b/pdfminer/cmapdb.py
@@ -118,6 +118,15 @@ def decode(self, code):
             return ()
 
 
+class IdentityCMapByte(IdentityCMap):
+    def decode(self, code):
+        n = len(code)
+        if n:
+            return struct.unpack('>%dB' % n, code)
+        else:
+            return ()
+
+
 class UnicodeMap(CMapBase):
     def __init__(self, **kwargs):
         CMapBase.__init__(self, **kwargs)
@@ -224,6 +233,10 @@ def get_cmap(klass, name):
             return IdentityCMap(WMode=0)
         elif name == 'Identity-V':
             return IdentityCMap(WMode=1)
+        elif name == 'OneByteIdentityH':
+            return IdentityCMapByte(WMode=0)
+        elif name == 'OneByteIdentityV':
+            return IdentityCMapByte(WMode=1)
         try:
             return klass._cmap_cache[name]
         except KeyError:

diff --git a/pdfminer/pdffont.py b/pdfminer/pdffont.py
@@ -946,7 +946,11 @@ def __init__(self, rsrcmgr, spec, strict=settings.STRICT):
         self.cidcoding = '%s-%s' % (resolve1(self.cidsysteminfo.get('Registry', b'unknown')).decode("latin1"),
                                     resolve1(self.cidsysteminfo.get('Ordering', b'unknown')).decode("latin1"))
         try:
-            name = literal_name(spec['Encoding'])
+            spec_encoding = spec['Encoding']
+            if hasattr(spec_encoding, 'name'):
+                name = literal_name(spec['Encoding'])
+            else:
+                name = literal_name(spec_encoding['CMapName'])
         except KeyError:
             if strict:
                 raise PDFFontError('Encoding is unspecified')

diff --git a/samples/Makefile b/samples/Makefile
@@ -18,7 +18,8 @@ NONFREE= \
 	nonfree/i1040nr \
 	nonfree/kampo \
 	nonfree/naacl06-shinyama \
-	nonfree/nlp2004slides
+	nonfree/nlp2004slides \
+	nonfree/one-byte-identity
 
 TESTS=$(FREE) $(NONFREE)
 

diff --git a/samples/nonfree/one-byte-identity.html.ref b/samples/nonfree/one-byte-identity.html.ref
@@ -0,0 +1,8 @@
+<html><head>
+<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+</head><body>
+<span style="position:absolute; border: gray 1px solid; left:0px; top:50px; width:595px; height:841px;"></span>
+<div style="position:absolute; top:50px;"><a name="1">Page 1</a></div>
+<div style="position:absolute; border: textbox 1px solid; writing-mode:lr-tb; left:30px; top:161px; width:108px; height:76px;"><span style="font-family: VGOQEA+SegoeUI; font-size:76px">abc
+<br></span></div><div style="position:absolute; top:0px;">Page: <a href="#1">1</a></div>
+</body></html>
diff --git a/samples/nonfree/one-byte-identity.pdf b/samples/nonfree/one-byte-identity.pdf
diff --git a/samples/nonfree/one-byte-identity.txt.ref b/samples/nonfree/one-byte-identity.txt.ref
@@ -0,0 +1,3 @@
+abc
+
+
diff --git a/samples/nonfree/one-byte-identity.xml.ref b/samples/nonfree/one-byte-identity.xml.ref
@@ -0,0 +1,17 @@
+<?xml version="1.0" encoding="utf-8" ?>
+<pages>
+<page id="1" bbox="0.000,0.000,595.280,841.890" rotate="0">
+<textbox id="0" bbox="30.000,653.550,138.990,730.270">
+<textline bbox="30.000,653.550,138.990,730.270">
+<text font="VGOQEA+SegoeUI" bbox="30.000,653.550,65.630,730.270" colourspace="CalGray" ncolour="None" size="76.720">a</text>
+<text font="VGOQEA+SegoeUI" bbox="65.560,653.550,106.720,730.270" colourspace="CalGray" ncolour="None" size="76.720">b</text>
+<text font="VGOQEA+SegoeUI" bbox="106.650,653.550,138.990,730.270" colourspace="CalGray" ncolour="None" size="76.720">c</text>
+<text>
+</text>
+</textline>
+</textbox>
+<layout>
+<textbox id="0" bbox="30.000,653.550,138.990,730.270" />
+</layout>
+</page>
+</pages>