diff --git a/CliClient/locales/fr_FR.po b/CliClient/locales/fr_FR.po
index a0ea0606b..05cbce08c 100644
--- a/CliClient/locales/fr_FR.po
+++ b/CliClient/locales/fr_FR.po
@@ -14,6 +14,8 @@ msgstr ""
"Content-Type: text/plain; charset=UTF-8\n"
"Content-Transfer-Encoding: 8bit\n"
"X-Generator: Poedit 2.0.3\n"
+"POT-Creation-Date: \n"
+"PO-Revision-Date: \n"
msgid "To delete a tag, untag the associated notes."
msgstr "Pour supprimer une vignette, enlever là des notes associées."
@@ -766,7 +768,7 @@ msgid ""
"how the notes or notebooks were originally encrypted."
msgstr ""
"Note : seule une clef maître va être utilisée pour le cryptage (celle "
-"marquée comme \"actif\" ci-dessus). N'importe quel clef peut-être utilisée "
+"marquée comme \"actif\" ci-dessus). N'importe quelle clef peut être utilisée "
"pour le décryptage, selon la façon dont les notes ou carnets étaient cryptés "
"à l'origine."
diff --git a/ReactNativeClient/lib/HtmlToMarkdownParser.js b/ReactNativeClient/lib/HtmlToMarkdownParser.js
new file mode 100644
index 000000000..075c45f92
--- /dev/null
+++ b/ReactNativeClient/lib/HtmlToMarkdownParser.js
@@ -0,0 +1,16 @@
+const { enexXmlToMd } = require('lib/import-enex-md-gen.js');
+const stringToStream = require('string-to-stream')
+
+class HtmlToMarkdownParser {
+
+ async parse(html, options = {}) {
+ if (!options.baseUrl) options.baseUrl = '';
+
+ const contentStream = stringToStream(html);
+ const markdown = await enexXmlToMd(contentStream, [], options);
+ return markdown;
+ }
+
+}
+
+module.exports = HtmlToMarkdownParser;
\ No newline at end of file
diff --git a/ReactNativeClient/lib/import-enex-md-gen.js b/ReactNativeClient/lib/import-enex-md-gen.js
index 03bb0dd92..861a0f147 100644
--- a/ReactNativeClient/lib/import-enex-md-gen.js
+++ b/ReactNativeClient/lib/import-enex-md-gen.js
@@ -403,7 +403,7 @@ function addResourceTag(lines, resource, alt = "") {
function isBlockTag(n) {
- return ["div", "p", "dl", "dd", 'dt', "center", 'address'].indexOf(n) >= 0;
+ return ["div", "p", "dl", "dd", 'dt', "center", 'address', 'form', 'input', 'section', 'nav', 'header', 'article', 'textarea', 'footer', 'fieldset'].indexOf(n) >= 0;
}
function isStrongTag(n) {
@@ -423,7 +423,7 @@ function isAnchor(n) {
}
function isIgnoredEndTag(n) {
- return ["en-note", "en-todo", "span", "body", "html", "font", "br", 'hr', 'tbody', 'sup', 'img', 'abbr', 'cite', 'thead', 'small', 'tt', 'sub', 'colgroup', 'col', 'ins', 'caption', 'var', 'map', 'area'].indexOf(n) >= 0;
+ return ["en-note", "en-todo", "span", "body", "html", "font", "br", 'hr', 'tbody', 'sup', 'img', 'abbr', 'cite', 'thead', 'small', 'tt', 'sub', 'colgroup', 'col', 'ins', 'caption', 'var', 'map', 'area', 'label', 'legend'].indexOf(n) >= 0;
}
function isListTag(n) {
@@ -432,7 +432,12 @@ function isListTag(n) {
// Elements that don't require any special treatment beside adding a newline character
function isNewLineOnlyEndTag(n) {
- return ["div", "p", "li", "h1", "h2", "h3", "h4", "h5", 'h6', "dl", "dd", 'dt', "center", 'address'].indexOf(n) >= 0;
+ return ["div", "p", "li", "h1", "h2", "h3", "h4", "h5", 'h6', "dl", "dd", 'dt', "center", 'address', 'form', 'input', 'section', 'nav', 'header', 'article', 'textarea', 'footer', 'fieldset'].indexOf(n) >= 0;
+}
+
+// Tags that must be ignored - both the tag and its content.
+function isIgnoredContentTag(n) {
+ return ['script', 'style', 'iframe', 'select', 'option', 'button', 'video', 'source'].indexOf(n) >= 0
}
function isCodeTag(n) {
@@ -480,7 +485,36 @@ function attributeToLowerCase(node) {
return output;
}
-function enexXmlToMdArray(stream, resources) {
+function urlWithoutPath(url) {
+ const parsed = require('url').parse(url, true);
+ return parsed.protocol + '//' + parsed.host;
+}
+
+function urlProtocol(url) {
+ const parsed = require('url').parse(url, true);
+ return parsed.protocol;
+}
+
+const schemeRegex = /[a-zA-Z0-9\+\-\.]+:\/\//
+// Make sure baseUrl doesn't end with a slash
+function prependBaseUrl(url, baseUrl) {
+ if (!url) url = '';
+ if (!baseUrl) return url;
+ const matches = schemeRegex.exec(url);
+ if (matches) return url; // Don't prepend the base URL if the URL already has a scheme
+
+ if (url.length >= 2 && url.indexOf('//') === 0) { // If it starts with // it's a protcol-relative URL
+ return urlProtocol(baseUrl) + url;
+ } else if (url && url[0] === '/') { // If it starts with a slash, it's an absolute URL so it should be relative to the domain (and not to the full baseUrl)
+ return urlWithoutPath(baseUrl) + url;
+ } else {
+ return baseUrl + '/' + url;
+ }
+}
+
+function enexXmlToMdArray(stream, resources, options = {}) {
+ if (options.baseUrl) options.baseUrl = options.baseUrl.replace(/[\/]+$/, '');
+
let remainingResources = resources.slice();
const removeRemainingResource = (id) => {
@@ -500,11 +534,12 @@ function enexXmlToMdArray(stream, resources) {
inCodeblock: 0,
lists: [],
anchorAttributes: [],
+ ignoreContents: [],
};
- let options = {};
+ let saxStreamOptions = {};
let strict = false;
- var saxStream = require('sax').createStream(strict, options)
+ var saxStream = require('sax').createStream(strict, saxStreamOptions)
let section = {
type: 'text',
@@ -518,6 +553,7 @@ function enexXmlToMdArray(stream, resources) {
})
saxStream.on('text', function(text) {
+ if (state.ignoreContents.length) return;
if (['table', 'tr', 'tbody'].indexOf(section.type) >= 0) return;
section.lines = collapseWhiteSpaceAndAppend(section.lines, state, text);
})
@@ -543,6 +579,8 @@ function enexXmlToMdArray(stream, resources) {
// Start of note
} else if (isBlockTag(n)) {
section.lines.push(BLOCK_OPEN);
+ } else if (isIgnoredContentTag(n)) {
+ state.ignoreContents.push(true);
} else if (n == 'table') {
let newSection = {
type: 'table',
@@ -590,7 +628,7 @@ function enexXmlToMdArray(stream, resources) {
} else if (n == 'li') {
section.lines.push(BLOCK_OPEN);
if (!state.lists.length) {
- reject("Found tag without being inside a list"); // TODO: could be a warning, but nothing to handle warnings at the moment
+ console.warn("Found tag without being inside a list");
return;
}
@@ -610,11 +648,10 @@ function enexXmlToMdArray(stream, resources) {
} else if (n == 'q') {
section.lines.push('"');
} else if (n == 'img') {
- // TODO: TEST IMAGE
if (nodeAttributes.src) { // Many (most?) img tags don't have no source associated, especially when they were imported from HTML
let s = '';
+ s += '](' + prependBaseUrl(nodeAttributes.src, options.baseUrl) + ')';
section.lines.push(s);
}
} else if (isAnchor(n)) {
@@ -746,7 +783,7 @@ function enexXmlToMdArray(stream, resources) {
section.lines.push(MONOSPACE_OPEN);
}
}
- } else if (["span", "font", 'sup', 'cite', 'abbr', 'small', 'tt', 'sub', 'colgroup', 'col', 'ins', 'caption', 'var', 'map', 'area'].indexOf(n) >= 0) {
+ } else if (["span", "font", 'sup', 'cite', 'abbr', 'small', 'tt', 'sub', 'colgroup', 'col', 'ins', 'caption', 'var', 'map', 'area', 'label', 'legend'].indexOf(n) >= 0) {
// Inline tags that can be ignored in Markdown
} else {
console.warn("Unsupported start tag: " + n);
@@ -771,6 +808,8 @@ function enexXmlToMdArray(stream, resources) {
// End of note
} else if (isNewLineOnlyEndTag(n)) {
section.lines.push(BLOCK_CLOSE);
+ } else if (isIgnoredContentTag(n)) {
+ state.ignoreContents.pop();
} else if (n == 'td' || n == 'th') {
if (section && section.parent) section = section.parent;
} else if (n == 'tr') {
@@ -806,6 +845,7 @@ function enexXmlToMdArray(stream, resources) {
} else if (isAnchor(n)) {
let attributes = state.anchorAttributes.pop();
let url = attributes && attributes.href ? attributes.href : '';
+ url = prependBaseUrl(url, options.baseUrl);
if (section.lines.length < 1) throw new Error('Invalid anchor tag closing'); // Sanity check, but normally not possible
@@ -813,10 +853,47 @@ function enexXmlToMdArray(stream, resources) {
// put the URL as is (don't wrap it in [](url)). The markdown parser, using
// GitHub flavour, will turn this URL into a link. This is to generate slightly
// cleaner markdown.
- let previous = section.lines[section.lines.length - 1];
+
+ // Need to loop on the previous tags so as to skip the special ones, which are not relevant for the below algorithm.
+ let previous = null;
+ for (let i = section.lines.length - 1; i >= 0; i--) {
+ previous = section.lines[i];
+ if ([BLOCK_OPEN, BLOCK_CLOSE, NEWLINE, NEWLINE_MERGED, SPACE].indexOf(previous) >= 0) {
+ continue;
+ } else {
+ break;
+ }
+ }
+
if (previous == '[') {
- section.lines.pop();
- section.lines.push(url);
+ // We have a link that had some content but, after parsing, nothing is left. The content was most likely
+ // something that shows up via CSS and which we cannot support. For example:
+ //
+ //
+ //
+ //
+ //
+ // In the case above the arrow is displayed via CSS.
+ // It is useless to display the full URL since often it is not relevant for a note (for example
+ // it's interactive bits) and it's not user-generated content such as a URL that would appear in a comment.
+ // So in this case, we still want to preserve the information but display it in a discreet way as a simple [L].
+
+ // Need to pop everything inside the current [] because it can only be special chars that we don't want (they would create uncessary newlines)
+ for (let i = section.lines.length - 1; i >= 0; i--) {
+ if (section.lines[i] !== '[') {
+ section.lines.pop();
+ } else {
+ break;
+ }
+ }
+
+ if (!url) {
+ // If there's no URL and no content, pop the [ and don't save any content.
+ section.lines.pop();
+ } else {
+ section.lines.push('(L)');
+ section.lines.push('](' + url + ')');
+ }
} else if (!previous || previous == url) {
section.lines.pop();
section.lines.pop();
@@ -995,8 +1072,8 @@ function drawTable(table) {
return flatRender ? lines : lines.join('<<<<:D>>>>' + NEWLINE + '<<<<:D>>>>').split('<<<<:D>>>>');
}
-async function enexXmlToMd(stream, resources) {
- let result = await enexXmlToMdArray(stream, resources);
+async function enexXmlToMd(stream, resources, options = {}) {
+ let result = await enexXmlToMdArray(stream, resources, options);
let mdLines = [];