From c343f1a90bc35d745de673de5ff771ddbe60be54 Mon Sep 17 00:00:00 2001
From: Jesse Rosenthal <jrosenthal@jhu.edu>
Date: Wed, 25 Jun 2014 08:10:19 -0400
Subject: Docx Reader: Add change types
Insertion and deletion. Dates are just strings for now.
---
src/Text/Pandoc/Readers/Docx/Parse.hs | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
index 18200bcf9..1cb5fe2e3 100644
--- a/src/Text/Pandoc/Readers/Docx/Parse.hs
+++ b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -455,6 +455,8 @@ elemToCell ns element
elemToCell _ _ = Nothing
data ParPart = PlainRun Run
+ | Insertion ChangeId Author ChangeDate [Run]
+ | Deletion ChangeId Author ChangeDate [Run]
| BookMark BookMarkId Anchor
| InternalHyperLink Anchor [Run]
| ExternalHyperLink RelId [Run]
@@ -604,4 +606,6 @@ type Target = String
type Anchor = String
type BookMarkId = String
type RelId = String
-
+type ChangeId = String
+type Author = String
+type ChangeDate = String
--
cgit v1.2.3
From 38e1d3e95b8240eeb35db0a1a56e308cfb4835e4 Mon Sep 17 00:00:00 2001
From: Jesse Rosenthal <jrosenthal@jhu.edu>
Date: Wed, 25 Jun 2014 10:32:48 -0400
Subject: Docx reader: Parse Insertions and Deletions.
This is just for the Parse module, reading it into the Docx format. It
still has to be translated into pandoc.
---
src/Text/Pandoc/Readers/Docx/Parse.hs | 18 +++++++++++++++++-
1 file changed, 17 insertions(+), 1 deletion(-)
diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
index 1cb5fe2e3..c76ef7511 100644
--- a/src/Text/Pandoc/Readers/Docx/Parse.hs
+++ b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -541,7 +541,7 @@ elemToRun _ _ = Nothing
elemToRunElem :: NameSpaces -> Element -> Maybe RunElem
elemToRunElem ns element
- | qName (elName element) == "t" &&
+ | (qName (elName element) == "t" || qName (elName element) == "delText") &&
qURI (elName element) == (lookup "w" ns) =
Just $ TextRun (strContent element)
| qName (elName element) == "br" &&
@@ -581,6 +581,22 @@ elemToParPart ns element
Nothing -> do
r <- elemToRun ns element
return $ PlainRun r
+elemToParPart ns element
+ | qName (elName element) == "ins" &&
+ qURI (elName element) == (lookup "w" ns) = do
+ cId <- findAttr (QName "id" (lookup "w" ns) (Just "w")) element
+ cAuthor <- findAttr (QName "author" (lookup "w" ns) (Just "w")) element
+ cDate <- findAttr (QName "date" (lookup "w" ns) (Just "w")) element
+ let runs = mapMaybe (elemToRun ns) (elChildren element)
+ return $ Insertion cId cAuthor cDate runs
+elemToParPart ns element
+ | qName (elName element) == "del" &&
+ qURI (elName element) == (lookup "w" ns) = do
+ cId <- findAttr (QName "id" (lookup "w" ns) (Just "w")) element
+ cAuthor <- findAttr (QName "author" (lookup "w" ns) (Just "w")) element
+ cDate <- findAttr (QName "date" (lookup "w" ns) (Just "w")) element
+ let runs = mapMaybe (elemToRun ns) (elChildren element)
+ return $ Deletion cId cAuthor cDate runs
elemToParPart ns element
| qName (elName element) == "bookmarkStart" &&
qURI (elName element) == (lookup "w" ns) = do
--
cgit v1.2.3
From ed44e4ca8c8f3d3c4c7ac65b98f16732c8173b88 Mon Sep 17 00:00:00 2001
From: Jesse Rosenthal <jrosenthal@jhu.edu>
Date: Wed, 25 Jun 2014 10:38:01 -0400
Subject: Docx reader: Add rudimentary track changes support.
This will only read the insertions, and ignore the deletions.
---
src/Text/Pandoc/Readers/Docx.hs | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/Text/Pandoc/Readers/Docx.hs b/src/Text/Pandoc/Readers/Docx.hs
index b787ca9fb..130e2a1e2 100644
--- a/src/Text/Pandoc/Readers/Docx.hs
+++ b/src/Text/Pandoc/Readers/Docx.hs
@@ -234,6 +234,9 @@ runToInlines opts docx@(Docx _ notes _ _ _) (Endnote fnId) =
parPartToInlines :: ReaderOptions -> Docx -> ParPart -> [Inline]
parPartToInlines opts docx (PlainRun r) = runToInlines opts docx r
+parPartToInlines opts docx (Insertion _ _ _ runs) =
+ concatMap (runToInlines opts docx) runs
+parPartToInlines _ _ (Deletion _ _ _ _) = []
parPartToInlines _ _ (BookMark _ anchor) | anchor `elem` dummyAnchors = []
parPartToInlines _ _ (BookMark _ anchor) = [Span (anchor, ["anchor"], []) []]
parPartToInlines _ (Docx _ _ _ rels _) (Drawing relid) =
--
cgit v1.2.3
From 9614ddfedc18cccbf9fbe1a23fae200c7e67d72d Mon Sep 17 00:00:00 2001
From: Jesse Rosenthal <jrosenthal@jhu.edu>
Date: Wed, 25 Jun 2014 11:00:15 -0400
Subject: Docx reader: Remove unnecessary filter in Parse.
mapMaybe does the filtering for us.
---
src/Text/Pandoc/Readers/Docx/Parse.hs | 9 ++-------
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
index c76ef7511..dbbd65681 100644
--- a/src/Text/Pandoc/Readers/Docx/Parse.hs
+++ b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -281,10 +281,6 @@ elemToBody ns element | qName (elName element) == "body" && qURI (elName element
$ map (elemToBodyPart ns) $ filterChildrenName (isParOrTbl ns) element
elemToBody _ _ = Nothing
-isRunOrLinkOrBookmark :: NameSpaces -> QName -> Bool
-isRunOrLinkOrBookmark ns q = qName q `elem` ["r", "hyperlink", "bookmarkStart"] &&
- qURI q == (lookup "w" ns)
-
elemToNumInfo :: NameSpaces -> Element -> Maybe (String, String)
elemToNumInfo ns element
| qName (elName element) == "p" &&
@@ -319,9 +315,8 @@ elemToBodyPart ns element
| qName (elName element) == "p" &&
qURI (elName element) == (lookup "w" ns) =
let parstyle = elemToParagraphStyle ns element
- parparts = mapMaybe id
- $ map (elemToParPart ns)
- $ filterChildrenName (isRunOrLinkOrBookmark ns) element
+ parparts = mapMaybe (elemToParPart ns)
+ $ elChildren element
in
case elemToNumInfo ns element of
Just (numId, lvl) -> Just $ ListItem parstyle numId lvl parparts
--
cgit v1.2.3
From a2b6ab847cb1c997c6ae7b8ed36f543a7ed90ecd Mon Sep 17 00:00:00 2001
From: Jesse Rosenthal <jrosenthal@jhu.edu>
Date: Wed, 25 Jun 2014 11:09:28 -0400
Subject: Docx reader: Add tests for basic track changes
This is what seems like the sensible default: read in insertions, and
ignore deletions. In the future, it would be good if options were
available for either taking in deletions or keeping both in some
scriptable format.
---
tests/Tests/Readers/Docx.hs | 10 ++++++++++
tests/docx.track_changes_deletion.docx | Bin 0 -> 13350 bytes
tests/docx.track_changes_deletion_only_ins.native | 1 +
tests/docx.track_changes_insertion.docx | Bin 0 -> 12956 bytes
tests/docx.track_changes_insertion_only_ins.native | 1 +
5 files changed, 12 insertions(+)
create mode 100644 tests/docx.track_changes_deletion.docx
create mode 100644 tests/docx.track_changes_deletion_only_ins.native
create mode 100644 tests/docx.track_changes_insertion.docx
create mode 100644 tests/docx.track_changes_insertion_only_ins.native
diff --git a/tests/Tests/Readers/Docx.hs b/tests/Tests/Readers/Docx.hs
index 4d062bbc0..f34e123ed 100644
--- a/tests/Tests/Readers/Docx.hs
+++ b/tests/Tests/Readers/Docx.hs
@@ -122,5 +122,15 @@ tests = [ testGroup "inlines"
"docx.codeblock.native"
]
+ , testGroup "track changes"
+ [ testCompare
+ "insert insertion (insertions only)"
+ "docx.track_changes_insertion.docx"
+ "docx.track_changes_insertion_only_ins.native"
+ , testCompare
+ "skip deletion (insertions only)"
+ "docx.track_changes_deletion.docx"
+ "docx.track_changes_deletion_only_ins.native"
+ ]
]
diff --git a/tests/docx.track_changes_deletion.docx b/tests/docx.track_changes_deletion.docx
new file mode 100644
index 000000000..5cfdbeed8
Binary files /dev/null and b/tests/docx.track_changes_deletion.docx differ
diff --git a/tests/docx.track_changes_deletion_only_ins.native b/tests/docx.track_changes_deletion_only_ins.native
new file mode 100644
index 000000000..205c67810
--- /dev/null
+++ b/tests/docx.track_changes_deletion_only_ins.native
@@ -0,0 +1 @@
+[Para [Str "This",Space,Str "is",Space,Str "a",Space,Str "text",Space,Str "with",Space,Str "a",Space,Str "deletion."]]
diff --git a/tests/docx.track_changes_insertion.docx b/tests/docx.track_changes_insertion.docx
new file mode 100644
index 000000000..fbdc9003e
Binary files /dev/null and b/tests/docx.track_changes_insertion.docx differ
diff --git a/tests/docx.track_changes_insertion_only_ins.native b/tests/docx.track_changes_insertion_only_ins.native
new file mode 100644
index 000000000..ca2e46df0
--- /dev/null
+++ b/tests/docx.track_changes_insertion_only_ins.native
@@ -0,0 +1 @@
+[Para [Str "This",Space,Str "is",Space,Str "a",Space,Str "text",Space,Str "with",Space,Str "two",Space,Str "exciting",Space,Str "insertions."]]
--
cgit v1.2.3