element-web/src/editor/serialize.ts

212 lines
8.6 KiB
TypeScript
Raw Normal View History

2019-05-22 17:16:32 +03:00
/*
Copyright 2019 New Vector Ltd
Copyright 2019, 2020 The Matrix.org Foundation C.I.C.
2019-05-22 17:16:32 +03:00
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
import Markdown from '../Markdown';
import {makeGenericPermalink} from "../utils/permalinks/Permalinks";
import EditorModel from "./model";
import { AllHtmlEntities } from 'html-entities';
2020-10-10 18:32:49 +03:00
import SettingsStore from '../settings/SettingsStore';
import SdkConfig from '../SdkConfig';
import cheerio from 'cheerio';
export function mdSerialize(model: EditorModel) {
return model.parts.reduce((html, part) => {
switch (part.type) {
case "newline":
return html + "\n";
case "plain":
2019-08-21 16:27:50 +03:00
case "command":
case "pill-candidate":
case "at-room-pill":
return html + part.text;
case "room-pill":
// Here we use the resourceId for compatibility with non-rich text clients
// See https://github.com/vector-im/element-web/issues/16660
return html +
`[${part.resourceId.replace(/[[\\\]]/g, c => "\\" + c)}](${makeGenericPermalink(part.resourceId)})`;
case "user-pill":
return html +
`[${part.text.replace(/[[\\\]]/g, c => "\\" + c)}](${makeGenericPermalink(part.resourceId)})`;
}
}, "");
}
2021-01-29 15:05:49 +03:00
export function htmlSerializeIfNeeded(model: EditorModel, {forceHTML = false} = {}) {
let md = mdSerialize(model);
// copy of raw input to remove unwanted math later
const orig = md;
2020-10-10 18:32:49 +03:00
if (SettingsStore.getValue("feature_latex_maths")) {
2021-04-01 13:09:51 +03:00
const patternNames = ['tex', 'latex'];
const patternTypes = ['display', 'inline'];
const patternDefaults = {
"tex": {
// detect math with tex delimiters, inline: $...$, display $$...$$
// preferably use negative lookbehinds, not supported in all major browsers:
// const displayPattern = "^(?<!\\\\)\\$\\$(?![ \\t])(([^$]|\\\\\\$)+?)\\$\\$$";
// const inlinePattern = "(?:^|\\s)(?<!\\\\)\\$(?!\\s)(([^$]|\\\\\\$)+?)(?<!\\\\|\\s)\\$";
// conditions for display math detection $$...$$:
// - pattern starts at beginning of line or is not prefixed with backslash or dollar
// - left delimiter ($$) is not escaped by backslash
"display": "(^|[^\\\\$])\\$\\$(([^$]|\\\\\\$)+?)\\$\\$",
// conditions for inline math detection $...$:
// - pattern starts at beginning of line, follows whitespace character or punctuation
// - pattern is on a single line
// - left and right delimiters ($) are not escaped by backslashes
// - left delimiter is not followed by whitespace character
// - right delimiter is not prefixed with whitespace character
"inline":
"(^|\\s|[.,!?:;])(?!\\\\)\\$(?!\\s)(([^$\\n]|\\\\\\$)*([^\\\\\\s\\$]|\\\\\\$)(?:\\\\\\$)?)\\$",
},
"latex": {
// detect math with latex delimiters, inline: \(...\), display \[...\]
// conditions for display math detection \[...\]:
// - pattern starts at beginning of line or is not prefixed with backslash
// - pattern is not empty
"display": "(^|[^\\\\])\\\\\\[(?!\\\\\\])(.*?)\\\\\\]",
// conditions for inline math detection \(...\):
// - pattern starts at beginning of line or is not prefixed with backslash
// - pattern is not empty
"inline": "(^|[^\\\\])\\\\\\((?!\\\\\\))(.*?)\\\\\\)",
},
};
patternNames.forEach(function(patternName) {
patternTypes.forEach(function(patternType) {
// get the regex replace pattern from config or use the default
const pattern = (((SdkConfig.get()["latex_maths_delims"] ||
{})[patternType] || {})["pattern"] || {})[patternName] ||
2021-04-01 13:09:51 +03:00
patternDefaults[patternName][patternType];
md = md.replace(RegExp(pattern, "gms"), function(m, p1, p2) {
const p2e = AllHtmlEntities.encode(p2);
switch (patternType) {
case "display":
return `${p1}<div data-mx-maths="${p2e}">\n\n</div>\n\n`;
case "inline":
return `${p1}<span data-mx-maths="${p2e}"></span>`;
}
});
});
2021-01-29 15:05:49 +03:00
});
// make sure div tags always start on a new line, otherwise it will confuse
// the markdown parser
md = md.replace(/(.)<div/g, function(m, p1) { return `${p1}\n<div`; });
}
const parser = new Markdown(md);
2019-07-08 17:55:56 +03:00
if (!parser.isPlainText() || forceHTML) {
// feed Markdown output to HTML parser
const phtml = cheerio.load(parser.toHTML(),
{ _useHtmlParser2: true, decodeEntities: false });
if (SettingsStore.getValue("feature_latex_maths")) {
// original Markdown without LaTeX replacements
const parserOrig = new Markdown(orig);
const phtmlOrig = cheerio.load(parserOrig.toHTML(),
{ _useHtmlParser2: true, decodeEntities: false });
// since maths delimiters are handled before Markdown,
// code blocks could contain mangled content.
// replace code blocks with original content
2021-04-07 18:22:30 +03:00
phtml('code').contents().each(function(i) {
const origData = phtmlOrig('code').contents()[i].data;
phtml('code').contents()[i].data = origData;
});
// add fallback output for latex math, which should not be interpreted as markdown
phtml('div, span').each(function(i, e) {
const tex = phtml(e).attr('data-mx-maths')
if (tex) {
phtml(e).html(`<code>${tex}</code>`)
}
});
}
return phtml.html();
}
// ensure removal of escape backslashes in non-Markdown messages
if (md.indexOf("\\") > -1) {
return parser.toPlaintext();
}
}
export function textSerialize(model: EditorModel) {
return model.parts.reduce((text, part) => {
switch (part.type) {
case "newline":
return text + "\n";
case "plain":
2019-08-21 16:27:50 +03:00
case "command":
case "pill-candidate":
case "at-room-pill":
return text + part.text;
case "room-pill":
// Here we use the resourceId for compatibility with non-rich text clients
// See https://github.com/vector-im/element-web/issues/16660
return text + `${part.resourceId}`;
case "user-pill":
return text + `${part.text}`;
}
}, "");
}
export function containsEmote(model: EditorModel) {
2020-06-16 16:06:42 +03:00
return startsWith(model, "/me ", false);
}
2020-06-16 16:06:42 +03:00
export function startsWith(model: EditorModel, prefix: string, caseSensitive = true) {
const firstPart = model.parts[0];
2019-08-21 16:27:50 +03:00
// part type will be "plain" while editing,
// and "command" while composing a message.
let text = firstPart && firstPart.text;
2020-06-16 16:06:42 +03:00
if (!caseSensitive) {
prefix = prefix.toLowerCase();
text = text.toLowerCase();
}
return firstPart && (firstPart.type === "plain" || firstPart.type === "command") && text.startsWith(prefix);
}
export function stripEmoteCommand(model: EditorModel) {
// trim "/me "
return stripPrefix(model, "/me ");
}
export function stripPrefix(model: EditorModel, prefix: string) {
model = model.clone();
model.removeText({index: 0, offset: 0}, prefix.length);
return model;
}
export function unescapeMessage(model: EditorModel) {
const {parts} = model;
if (parts.length) {
const firstPart = parts[0];
// only unescape \/ to / at start of editor
if (firstPart.type === "plain" && firstPart.text.startsWith("\\/")) {
model = model.clone();
model.removeText({index: 0, offset: 0}, 1);
}
}
return model;
}