Skip to content

Commit 448d7ea

Browse files
K-Kumar-01K-Kumar-01
authored
fix: rewrite transformer(CiceroMarkToOOXML) logic to handle nesting - #397 (#418)
* feat(markdown-docx): text and emphasis transformer Remove old transformer Rules: EMPHASIS, TEXT, TEXT_STYLES, TEXT_WRAPPER, PARAGRAPH_WRAPPER Tests: Check only for text-and-emphasis using if Signed-off-by: k-kumar-01 <[email protected]> * feat: heading transformer Logic to transform headings Rule: PARAGRAPH_PROPERTIES_RULE Hardcore check the test for heading using condition Signed-off-by: K-Kumar-01 <[email protected]> * feat: variable transformer Logic to transform variables Rules: Variable Rule Conditionally check for tests Signed-off-by: K-Kumar-01 <[email protected]> * feat: softbreak transformer Logic for softbreak transformation Rule: SOFTBREAK_RULE Conditionally exclude the test for strong(file:strong.json) Signed-off-by: K-Kumar-01 <[email protected]> * feat: strong transformer Logic for strong transformation Rule: STRONG_RULE Check for all tests Signed-off-by: K-Kumar-01 <[email protected]> * feat: add headingStyles and relationship specs Signed-off-by: K-Kumar-01 <[email protected]> * refactor(markdown-docx): coding practices Spread operator use inplace of push Define constants as per convention Add test for nesting Signed-off-by: K-Kumar-01 <[email protected]> * refactor: Heading Properties remove argument value from Rule remove the condition to insert runtime properties in transformer Signed-off-by: K-Kumar-01 <[email protected]>
1 parent 0bd57ec commit 448d7ea

File tree

4 files changed

+343
-127
lines changed

4 files changed

+343
-127
lines changed

‎packages/markdown-docx/src/CiceroMarkToOOXML/helpers.js

+168
Original file line numberDiff line numberDiff line change
@@ -42,6 +42,172 @@ function titleGenerator(title, type) {
4242
* @returns {string} OOXML wraped in docx headers
4343
*/
4444
function wrapAroundDefaultDocxTags(ooxml) {
45+
46+
const HEADING_STYLE_SPEC = `
47+
<pkg:part pkg:name="/word/styles.xml" pkg:contentType="application/vnd.openxmlformats-officedocument.wordprocessingml.styles+xml">
48+
<pkg:xmlData>
49+
<w:styles xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006" xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships" xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main" xmlns:w14="http://schemas.microsoft.com/office/word/2010/wordml" xmlns:w15="http://schemas.microsoft.com/office/word/2012/wordml" xmlns:w16cid="http://schemas.microsoft.com/office/word/2016/wordml/cid" xmlns:w16se="http://schemas.microsoft.com/office/word/2015/wordml/symex" mc:Ignorable="w14 w15 w16se w16cid">
50+
<w:docDefaults>
51+
<w:rPrDefault>
52+
<w:rPr>
53+
<w:rFonts w:asciiTheme="minorHAnsi" w:eastAsiaTheme="minorHAnsi" w:hAnsiTheme="minorHAnsi" w:cstheme="minorBidi"/>
54+
<w:sz w:val="22"/>
55+
<w:szCs w:val="22"/>
56+
<w:lang w:val="en-US" w:eastAsia="en-US" w:bidi="ar-SA"/>
57+
</w:rPr>
58+
</w:rPrDefault>
59+
<w:pPrDefault>
60+
<w:pPr>
61+
<w:spacing w:after="160" w:line="259" w:lineRule="auto"/>
62+
</w:pPr>
63+
</w:pPrDefault>
64+
</w:docDefaults>
65+
<w:latentStyles w:defLockedState="0" w:defUIPriority="99" w:defSemiHidden="0" w:defUnhideWhenUsed="0" w:defQFormat="0" w:count="377">
66+
<w:lsdException w:name="Normal" w:uiPriority="0" w:qFormat="1"/>
67+
<w:lsdException w:name="heading 1" w:uiPriority="9" w:qFormat="1"/>
68+
<w:lsdException w:name="heading 2" w:semiHidden="1" w:uiPriority="9" w:unhideWhenUsed="1" w:qFormat="1"/>
69+
<w:lsdException w:name="heading 3" w:semiHidden="1" w:uiPriority="9" w:unhideWhenUsed="1" w:qFormat="1"/>
70+
<w:lsdException w:name="heading 4" w:semiHidden="1" w:uiPriority="9" w:unhideWhenUsed="1" w:qFormat="1"/>
71+
<w:lsdException w:name="heading 5" w:semiHidden="1" w:uiPriority="9" w:unhideWhenUsed="1" w:qFormat="1"/>
72+
<w:lsdException w:name="heading 6" w:semiHidden="1" w:uiPriority="9" w:unhideWhenUsed="1" w:qFormat="1"/>
73+
</w:latentStyles>
74+
<w:style w:type="paragraph" w:styleId="Heading1">
75+
<w:name w:val="heading 1" />
76+
<w:basedOn w:val="Normal" />
77+
<w:next w:val="Normal" />
78+
<w:link w:val="Heading1Char" />
79+
<w:uiPriority w:val="9" />
80+
<w:qFormat />
81+
<w:pPr>
82+
<w:keepNext />
83+
<w:keepLines />
84+
<w:spacing w:before="240" w:after="0" />
85+
<w:outlineLvl w:val="0" />
86+
</w:pPr>
87+
<w:rPr>
88+
<w:rFonts w:asciiTheme="majorHAnsi" w:eastAsiaTheme="majorEastAsia" w:hAnsiTheme="majorHAnsi" w:cstheme="majorBidi" />
89+
<w:color w:val="2E74B5" w:themeColor="accent1" w:themeShade="BF" />
90+
<w:sz w:val="32" />
91+
<w:szCs w:val="32" />
92+
</w:rPr>
93+
</w:style>
94+
<w:style w:type="paragraph" w:styleId="Heading2">
95+
<w:name w:val="heading 2"/>
96+
<w:basedOn w:val="Normal"/>
97+
<w:next w:val="Normal"/>
98+
<w:link w:val="Heading2Char"/>
99+
<w:uiPriority w:val="9"/>
100+
<w:unhideWhenUsed/>
101+
<w:qFormat/>
102+
<w:pPr>
103+
<w:keepNext/>
104+
<w:keepLines/>
105+
<w:spacing w:before="40" w:after="0"/>
106+
<w:outlineLvl w:val="1"/>
107+
</w:pPr>
108+
<w:rPr>
109+
<w:rFonts w:asciiTheme="majorHAnsi" w:eastAsiaTheme="majorEastAsia" w:hAnsiTheme="majorHAnsi" w:cstheme="majorBidi"/>
110+
<w:color w:val="2E74B5" w:themeColor="accent1" w:themeShade="BF"/>
111+
<w:sz w:val="26"/>
112+
<w:szCs w:val="26"/>
113+
</w:rPr>
114+
</w:style>
115+
<w:style w:type="paragraph" w:styleId="Heading3">
116+
<w:name w:val="heading 3" />
117+
<w:basedOn w:val="Normal" />
118+
<w:next w:val="Normal" />
119+
<w:link w:val="Heading3Char" />
120+
<w:uiPriority w:val="9" />
121+
<w:unhideWhenUsed />
122+
<w:qFormat />
123+
<w:pPr>
124+
<w:keepNext />
125+
<w:keepLines />
126+
<w:spacing w:before="40" w:after="0" />
127+
<w:outlineLvl w:val="2" />
128+
</w:pPr>
129+
<w:rPr>
130+
<w:rFonts w:asciiTheme="majorHAnsi" w:eastAsiaTheme="majorEastAsia" w:hAnsiTheme="majorHAnsi" w:cstheme="majorBidi" />
131+
<w:color w:val="1F4D78" w:themeColor="accent1" w:themeShade="7F" />
132+
<w:sz w:val="24" />
133+
<w:szCs w:val="24" />
134+
</w:rPr>
135+
</w:style>
136+
<w:style w:type="paragraph" w:styleId="Heading4">
137+
<w:name w:val="heading 4" />
138+
<w:basedOn w:val="Normal" />
139+
<w:next w:val="Normal" />
140+
<w:link w:val="Heading4Char" />
141+
<w:uiPriority w:val="9" />
142+
<w:unhideWhenUsed />
143+
<w:qFormat />
144+
<w:pPr>
145+
<w:keepNext />
146+
<w:keepLines />
147+
<w:spacing w:before="40" w:after="0" />
148+
<w:outlineLvl w:val="3" />
149+
</w:pPr>
150+
<w:rPr>
151+
<w:rFonts w:asciiTheme="majorHAnsi" w:eastAsiaTheme="majorEastAsia" w:hAnsiTheme="majorHAnsi" w:cstheme="majorBidi" />
152+
<w:i />
153+
<w:iCs />
154+
<w:color w:val="2E74B5" w:themeColor="accent1" w:themeShade="BF" />
155+
</w:rPr>
156+
</w:style>
157+
<w:style w:type="paragraph" w:styleId="Heading5">
158+
<w:name w:val="heading 5" />
159+
<w:basedOn w:val="Normal" />
160+
<w:next w:val="Normal" />
161+
<w:link w:val="Heading5Char" />
162+
<w:uiPriority w:val="9" />
163+
<w:unhideWhenUsed />
164+
<w:qFormat />
165+
<w:pPr>
166+
<w:keepNext />
167+
<w:keepLines />
168+
<w:spacing w:before="40" w:after="0" />
169+
<w:outlineLvl w:val="4" />
170+
</w:pPr>
171+
<w:rPr>
172+
<w:rFonts w:asciiTheme="majorHAnsi" w:eastAsiaTheme="majorEastAsia" w:hAnsiTheme="majorHAnsi" w:cstheme="majorBidi" />
173+
<w:color w:val="2E74B5" w:themeColor="accent1" w:themeShade="BF" />
174+
</w:rPr>
175+
</w:style>
176+
<w:style w:type="paragraph" w:styleId="Heading6">
177+
<w:name w:val="heading 6" />
178+
<w:basedOn w:val="Normal" />
179+
<w:next w:val="Normal" />
180+
<w:link w:val="Heading6Char" />
181+
<w:uiPriority w:val="9" />
182+
<w:unhideWhenUsed />
183+
<w:qFormat />
184+
<w:pPr>
185+
<w:keepNext />
186+
<w:keepLines />
187+
<w:spacing w:before="40" w:after="0" />
188+
<w:outlineLvl w:val="5" />
189+
</w:pPr>
190+
<w:rPr>
191+
<w:rFonts w:asciiTheme="majorHAnsi" w:eastAsiaTheme="majorEastAsia" w:hAnsiTheme="majorHAnsi" w:cstheme="majorBidi" />
192+
<w:color w:val="1F4D78" w:themeColor="accent1" w:themeShade="7F" />
193+
</w:rPr>
194+
</w:style>
195+
</w:styles>
196+
</pkg:xmlData>
197+
</pkg:part>
198+
`;
199+
200+
const RELATIONSHIP_SPEC = `
201+
<pkg:part pkg:name="/word/_rels/document.xml.rels" pkg:contentType="application/vnd.openxmlformats-package.relationships+xml" pkg:padding="256">
202+
<pkg:xmlData>
203+
<Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
204+
<Relationship Id="rId1" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/numbering" Target="numbering.xml"/>
205+
<Relationship Id="rId2" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/styles" Target="styles.xml"/>
206+
</Relationships>
207+
</pkg:xmlData>
208+
</pkg:part>
209+
`;
210+
45211
ooxml = `<pkg:package
46212
xmlns:pkg="http://schemas.microsoft.com/office/2006/xmlPackage">
47213
<pkg:part pkg:name="/_rels/.rels" pkg:contentType="application/vnd.openxmlformats-package.relationships+xml" pkg:padding="512">
@@ -91,6 +257,8 @@ function wrapAroundDefaultDocxTags(ooxml) {
91257
</w:document>
92258
</pkg:xmlData>
93259
</pkg:part>
260+
${RELATIONSHIP_SPEC}
261+
${HEADING_STYLE_SPEC}
94262
</pkg:package>`;
95263

96264
return ooxml;

‎packages/markdown-docx/src/CiceroMarkToOOXML/index.js

+95-80
Original file line numberDiff line numberDiff line change
@@ -14,7 +14,17 @@
1414

1515
'use strict';
1616

17-
const { TEXT_RULE, EMPHASIS_RULE, HEADING_RULE, VARIABLE_RULE, SOFTBREAK_RULE, STRONG_RULE } = require('./rules');
17+
const {
18+
TEXT_RULE,
19+
EMPHASIS_RULE,
20+
PARAGRAPH_RULE,
21+
TEXT_STYLES_RULE,
22+
TEXT_WRAPPER_RULE,
23+
HEADING_PROPERTIES_RULE,
24+
VARIABLE_RULE,
25+
SOFTBREAK_RULE,
26+
STRONG_RULE,
27+
} = require('./rules');
1828
const { wrapAroundDefaultDocxTags } = require('./helpers');
1929

2030
const definedNodes = {
@@ -36,11 +46,15 @@ const definedNodes = {
3646
*/
3747
class CiceroMarkToOOXMLTransfomer {
3848
/**
39-
* Declares the OOXML and counter variable.
49+
* Declares the OOXML, counter, and tags variable.
4050
*/
4151
constructor() {
52+
// OOXML for given CiceroMark JSON
4253
this.globalOOXML = '';
54+
// Frequency of different variables in CiceroMark JSON
4355
this.counter = {};
56+
// OOXML tags for a given block node(heading, pargraph, etc.)
57+
this.tags = [];
4458
}
4559

4660
/**
@@ -63,90 +77,92 @@ class CiceroMarkToOOXMLTransfomer {
6377
}
6478

6579
/**
66-
* Gets the OOXML for the given node.
80+
* Traverses CiceroMark nodes in a DFS approach
6781
*
68-
* @param {object} node Description of node type
69-
* @param {object} parent Parent object for a node
70-
* @returns {string} OOXML for the given node
82+
* @param {object} node CiceroMark Node
83+
* @param {array} properties Properties to be applied on curent node
7184
*/
72-
getNodes(node, parent = null) {
73-
if (this.getClass(node) === definedNodes.variable) {
74-
const tag = node.name;
75-
const type = node.elementType;
76-
if (Object.prototype.hasOwnProperty.call(this.counter, tag)) {
77-
this.counter = {
78-
...this.counter,
79-
[tag]: {
80-
...this.counter[tag],
81-
count: ++this.counter[tag].count,
82-
},
83-
};
84-
} else {
85-
this.counter[tag] = {
86-
count: 1,
87-
type,
88-
};
89-
}
90-
const value = node.value;
91-
const title = `${tag.toUpperCase()[0]}${tag.substring(1)}${this.counter[tag].count}`;
92-
return VARIABLE_RULE(title, tag, value, type);
93-
}
85+
travserseNodes(node, properties = []) {
86+
if (this.getClass(node) === 'org.accordproject.commonmark.Document') {
87+
this.travserseNodes(node.nodes, properties);
88+
} else {
89+
for (let subNode of node) {
90+
if (this.getClass(subNode) === definedNodes.text) {
91+
let propertyTag = '';
92+
for (let property of properties) {
93+
if (property === definedNodes.emphasize) {
94+
propertyTag += EMPHASIS_RULE();
95+
} else if (property === definedNodes.strong) {
96+
propertyTag += STRONG_RULE();
97+
}
98+
}
99+
if (propertyTag) {
100+
propertyTag = TEXT_STYLES_RULE(propertyTag);
101+
}
94102

95-
if (this.getClass(node) === definedNodes.text) {
96-
if (parent !== null && parent.class === definedNodes.heading) {
97-
return HEADING_RULE(node.text, parent.level);
98-
}
99-
if (parent !== null && parent.class === definedNodes.emphasize) {
100-
return EMPHASIS_RULE(node.text);
101-
}
102-
if (parent !== null && parent.class === definedNodes.strong) {
103-
return STRONG_RULE(node.text);
104-
} else {
105-
return TEXT_RULE(node.text);
106-
}
107-
}
103+
let textValueTag = TEXT_RULE(subNode.text);
108104

109-
if (this.getClass(node) === definedNodes.softbreak) {
110-
return SOFTBREAK_RULE();
111-
}
105+
let tag = TEXT_WRAPPER_RULE(propertyTag, textValueTag);
106+
this.tags = [...this.tags, tag];
107+
} else if (this.getClass(subNode) === definedNodes.variable) {
108+
const tag = subNode.name;
109+
const type = subNode.elementType;
110+
if (Object.prototype.hasOwnProperty.call(this.counter, tag)) {
111+
this.counter = {
112+
...this.counter,
113+
[tag]: {
114+
...this.counter[tag],
115+
count: ++this.counter[tag].count,
116+
},
117+
};
118+
} else {
119+
this.counter[tag] = {
120+
count: 1,
121+
type,
122+
};
123+
}
124+
const value = subNode.value;
125+
const title = `${tag.toUpperCase()[0]}${tag.substring(1)}${this.counter[tag].count}`;
112126

113-
if (this.getClass(node) === definedNodes.strong) {
114-
let ooxml = '';
115-
node.nodes.forEach(subNode => {
116-
ooxml += this.getNodes(subNode, { class: node.$class });
117-
});
118-
return ooxml;
119-
}
127+
this.tags = [...this.tags, VARIABLE_RULE(title, tag, value, type)];
128+
} else if (this.getClass(subNode) === definedNodes.softbreak) {
129+
this.tags = [...this.tags, SOFTBREAK_RULE()];
130+
} else {
131+
if (subNode.nodes) {
132+
if (this.getClass(subNode) === definedNodes.paragraph) {
133+
this.travserseNodes(subNode.nodes, properties);
134+
let ooxml = '';
135+
for (let xmlTag of this.tags) {
136+
ooxml += xmlTag;
137+
}
138+
ooxml = PARAGRAPH_RULE(ooxml);
120139

121-
if (this.getClass(node) === definedNodes.emphasize) {
122-
let ooxml = '';
123-
node.nodes.forEach(subNode => {
124-
ooxml += this.getNodes(subNode, { class: node.$class });
125-
});
126-
return ooxml;
127-
}
140+
this.globalOOXML += ooxml;
141+
// Clear all the tags as all nodes of paragraph have been traversed.
142+
this.tags = [];
143+
} else if (this.getClass(subNode) === definedNodes.heading) {
144+
this.travserseNodes(subNode.nodes, properties);
145+
let ooxml = '';
146+
for (let xmlTag of this.tags) {
147+
let headingPropertiesTag = '';
148+
headingPropertiesTag = HEADING_PROPERTIES_RULE(subNode.level);
149+
ooxml += headingPropertiesTag;
150+
ooxml += xmlTag;
151+
}
128152

129-
if (this.getClass(node) === definedNodes.heading) {
130-
let ooxml = '';
131-
node.nodes.forEach(subNode => {
132-
ooxml += this.getNodes(subNode, { class: node.$class, level: node.level });
133-
});
134-
this.globalOOXML = `
135-
${this.globalOOXML}
136-
<w:p>
137-
${ooxml}
138-
</w:p>
139-
`;
140-
}
153+
// in DOCX heading is a paragraph with some styling tags present
154+
ooxml = PARAGRAPH_RULE(ooxml);
141155

142-
if (this.getClass(node) === definedNodes.paragraph) {
143-
let ooxml = '';
144-
node.nodes.forEach(subNode => {
145-
ooxml += this.getNodes(subNode);
146-
});
147-
this.globalOOXML = `${this.globalOOXML}<w:p>${ooxml}</w:p>`;
156+
this.globalOOXML += ooxml;
157+
this.tags = [];
158+
} else {
159+
let newProperties = [...properties, subNode.$class];
160+
this.travserseNodes(subNode.nodes, newProperties);
161+
}
162+
}
163+
}
164+
}
148165
}
149-
return '';
150166
}
151167

152168
/**
@@ -156,10 +172,9 @@ class CiceroMarkToOOXMLTransfomer {
156172
* @returns {string} Converted OOXML string i.e. CicecoMark->OOXML
157173
*/
158174
toOOXML(ciceromark) {
159-
ciceromark.nodes.forEach(node => {
160-
this.getNodes(node);
161-
});
175+
this.travserseNodes(ciceromark, []);
162176
this.globalOOXML = wrapAroundDefaultDocxTags(this.globalOOXML);
177+
163178
return this.globalOOXML;
164179
}
165180
}

0 commit comments

Comments
 (0)