@iebh/polyglot
Version:
IEBH-SRA tool to convert between different medical database search formats
612 lines (597 loc) • 22.6 kB
JavaScript
;
Object.defineProperty(exports, "__esModule", {
value: true
});
exports.parse = void 0;
var _global = _interopRequireDefault(require("./global.js"));
var _tools = _interopRequireDefault(require("./tools.js"));
var _lodash = _interopRequireDefault(require("lodash"));
var _fieldCodesParse = _interopRequireDefault(require("../data/fieldCodesParse.js"));
var _meshTranslationsParse = _interopRequireDefault(require("../data/meshTranslationsParse.js"));
function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; }
// Parsing Objects
// Escape all regular expression chars except for pipe
function escapeRegExp(string) {
return string.replace(/[.*+?^${}()[\]\\]/g, '\\$&'); // $& means the whole matched string
}
/**
* Parse a given string into a lexical object tree
* This tree can then be recompiled via each engines compile()
* @param {string} query The query string to compile. This can be multiline
* @param {Object} [options] Optional options to use when parsing
* @param {boolean} [options.groupLines=false] Wrap lines inside their own groups (only applies if multiple lines are present)
* @param {boolean} [options.groupLinesAlways=true] Group lines even if there is only one apparent line (i.e. enclose single line queries within brackets)
* @param {boolean} [options.transposeLines=true] Insert all line references where needed (e.g. `1 - 3/OR`)
* @param {boolean} [options.removeNumbering=true] Remove any number prefixes from lines - this is a classic copy/paste error from certain online search engines (e.g. `1. Term` -> `term`)
* @param {boolean} [options.preserveNewlines=true] Preserve newlines in the output as 'raw' tree nodes
* @return {array} Array representing the parsed tree nodes
*/
var parse = function parse(query, options) {
var settings = _lodash["default"].defaults(options, {
groupLines: false,
groupLinesAlways: false,
removeNumbering: false,
preserveNewlines: true,
transposeLines: true
});
_global["default"].variables.no_field_tag = []; // Empty array of offsets each time the query is parsed
var q = query + ''; // Clone query
var tree = {
nodes: []
}; // Tree is the full parsed tree
var branchStack = [tree]; // Stack for where we are within the tree (will get pushed when a new group is encountered)
var branch = tree; // Branch is the parent of leaf (branch always equals last element of branchStack)
var lastGroup; // Optional reference to the previously created group (used to pin things)
var leaf = branch.nodes; // Leaf is the currently active leaf node (usually branch.nodes)
var afterWhitespace = true; // Set to true when the current character is following whitespace, a newline or the very start of the query
var lineRefs = {}; // Object lookup for line references (usually something like `1. Foo`), only populated if `transposeLines` is true
// Operate in line-by-line mode? {{{
if (settings.transposeLines || settings.groupLines || settings.removeNumbering) {
var lines = q.split('\n');
// Remove numbering {{{
if (settings.removeNumbering) {
var match;
lines = lines.map(function (line) {
if (match = /^\s*\d+\.?\s(.*)$/.exec(line)) {
return match[1];
} else {
return line;
}
});
}
// }}}
// Group line content {{{
if (settings.groupLines && (settings.groupLinesAlways || lines.length > 1)) {
// Wrap lines provided they are not blank and are not just 'and', 'or', 'not' by themselves or a comment
lines = lines.map(function (line) {
return _lodash["default"].trim(line) && !/^\s*(and|or|not)\s*$/i.test(line) && !/^\s*#/.test(line) ? '(' + line + ')' : line;
});
}
// }}}
q = lines.join('\n'); // Join up lines again
}
// }}}
// Utility functions {{{
/**
* Trim previous leaf content if it has any text
* The leaf will be removed completely if it is now blank
*/
function trimLastLeaf() {
if (leaf && _lodash["default"].includes(['phrase', 'raw'], leaf.type) && / $/.test(leaf.content)) {
leaf.content = leaf.content.substr(0, leaf.content.length - 1);
if (!leaf.content) branch.nodes.pop();
}
}
;
/**
* End the previous line branch and create a new one
* this function is run every time a new raw node is inserted
*/
function newLine(currentNumber) {
lastGroup = branch;
branch = branchStack.pop();
leaf = branch.nodes;
var newGroup = {
type: 'line',
number: currentNumber,
isNumbered: false,
nodes: []
};
branch.nodes.push(newGroup);
branchStack.push(branch);
branch = newGroup;
leaf = branch.nodes;
}
;
// }}}
// Create a group for the first line
var newGroup = {
type: 'line',
number: 1,
isNumbered: false,
nodes: []
};
branch.nodes.push(newGroup);
branchStack.push(branch);
branch = newGroup;
leaf = branch.nodes;
var lineNumber = 1;
// Variable to store whether there is user entered line numbering
var userLineNumber = false;
// Variable to store byte offset of string at current point
var offset = 0;
// Create string of field codes seperated by pipe operator
var fieldCodes = escapeRegExp(Object.keys(_fieldCodesParse["default"]).join("|"));
var meshTranslations = escapeRegExp(Object.keys(_meshTranslationsParse["default"]).join("|"));
while (q.length) {
var cropString = true; // Whether to remove one charcater from the beginning of the string (set to false if the lexical match handles this behaviour itself)
var match;
if (/^\(/.test(q)) {
var newGroup = {
type: 'group',
nodes: []
};
branch.nodes.push(newGroup);
branchStack.push(branch);
branch = newGroup;
leaf = branch.nodes;
} else if (/^\)/.test(q)) {
lastGroup = branch;
if (branchStack.length > 0) {
branch = branchStack.pop();
} else {
// TODO: Code for popover message
// branch.msg = "Extra closing bracket removed after term"
}
leaf = branch.nodes;
} else if (match = /^([0-9]+)\s*[‐\-]\s*([0-9]+)(?:\/(AND|OR|NOT))/i.exec(q)) {
// 1-7/OR
branch.nodes.push({
type: 'ref',
ref: _lodash["default"].range(match[1], (match[2] + 1) / 10),
cond: match[3].toUpperCase(),
nodes: []
});
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^(AND|OR|NOT)(?:\/([0-9]+)\s*[‐\-]\s*([0-9]+))/i.exec(q)) {
// OR/1-7
branch.nodes.push({
type: 'ref',
ref: _lodash["default"].range(match[2], (match[3] + 1) / 10),
cond: match[1].toUpperCase(),
nodes: []
});
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^(OR)\/([0-9]+(?:,[0-9]+)*)/i.exec(q)) {
// OR/1,3
branch.nodes.push({
type: 'ref',
ref: match[2].split(','),
cond: match[1].toUpperCase(),
nodes: []
});
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^(AND)\/([0-9]+(?:,[0-9]+)*)/i.exec(q)) {
// AND/1,3
branch.nodes.push({
type: 'ref',
ref: match[2].split(',').map(function (num) {
return "#".concat(num.trim());
}),
cond: match[1].toUpperCase(),
nodes: []
});
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
// } else if (match = /^(OR)\/([0-9]+(?:,[0-9]+)*)/i.exec(q)) {
// // OR/11,14,..
// const refNumbers = match[2].split(',').map(num => `#${num.trim()}`).join(' OR ');
// branch.nodes.push({
// type: 'ref',
// ref: refNumbers,
// cond: 'OR',
// nodes: []
// });
// offset += match[0].length;
// q = q.substr(match[0].length);
// cropString = false;
} else if (match = /^(#?([0-9]+)) +(AND|OR|NOT)\s+/i.exec(q)) {
// 1 AND ...
if (leaf.type != "phrase") {
branch.nodes.push({
type: 'ref',
ref: [match[2]],
cond: '',
nodes: []
});
offset += match[1].length;
q = q.substr(match[1].length); // NOTE we only move by the digits, not the whole expression - so we can still handle the AND/OR correctly
cropString = false;
} else {
leaf.content += match[1];
offset += match[1].length - 1;
q = q.substr(match[1].length - 1);
}
} else if (match = /^((AND|OR|NOT) +#?([0-9]+))($(?![\r\n])|\s+|\))/i.exec(q)) {
// AND 2...
trimLastLeaf();
switch (match[2].toLowerCase()) {
case "and":
branch.nodes.push({
type: 'joinAnd'
});
break;
case "or":
branch.nodes.push({
type: 'joinOr'
});
break;
case "not":
branch.nodes.push({
type: 'joinNot'
});
break;
}
leaf = undefined;
cropString = false;
branch.nodes.push({
type: 'ref',
ref: [match[3]],
cond: '',
nodes: []
});
offset += match[1].length;
q = q.substr(match[1].length);
} else if (match = /^([0-9]+\.?)\s+/i.exec(q)) {
// 1 or 1. (Line number)
if (leaf.type != "phrase") {
lineNumber = parseInt(match[1], 10);
branch.number = lineNumber;
branch.isNumbered = true;
userLineNumber = true;
offset += match[0].length - 1;
q = q.substr(match[0].length - 1);
} else {
leaf.content += match[1];
offset += match[1].length - 1;
q = q.substr(match[1].length - 1);
}
} else if (afterWhitespace && (match = /^and\b/i.exec(q))) {
trimLastLeaf();
branch.nodes.push({
type: 'joinAnd'
});
leaf = undefined;
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (afterWhitespace && (match = /^or\b/i.exec(q))) {
trimLastLeaf();
branch.nodes.push({
type: 'joinOr'
});
leaf = undefined;
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (afterWhitespace && (match = /^not\b/i.exec(q))) {
trimLastLeaf();
branch.nodes.push({
type: 'joinNot'
});
leaf = undefined;
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (afterWhitespace && (match = /^(near\/|near|adj|n)(\d+)\b/i.exec(q))) {
trimLastLeaf();
if (match[2]) branch.nodes.push({
type: 'joinNear',
proximity: _lodash["default"].toNumber(match[2])
});else branch.nodes.push({
type: 'joinNear',
proximity: 1
});
leaf = undefined;
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (afterWhitespace && (match = /^(next\/|next|adj|w|w\/|pre\/|p\/)(\d+)?\b/i.exec(q))) {
trimLastLeaf();
if (match[2]) branch.nodes.push({
type: 'joinNext',
proximity: _lodash["default"].toNumber(match[2])
});else branch.nodes.push({
type: 'joinNext',
proximity: 1
});
leaf = undefined;
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
}
// MESHTRANSLATIONS {{{
else if (afterWhitespace && (match = new RegExp("^(".concat(meshTranslations, ")"), "i").exec(q.toLowerCase().replace(/"/g, '')))) {
branch.nodes.push({
type: 'meshTranslation',
field: _meshTranslationsParse["default"][match[1]]
});
match = /.+?[\.\[]\S+[\.\]]/.exec(q);
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
}
/// }}}
// MESH {{{
else if (match = /^\[(mesh terms|mesh|mh)(:NoExp|:no exp)?\]/i.exec(q)) {
// Mesh term - PubMed syntax
leaf.type = 'mesh';
leaf.field = match[2] ? 'Mesh search (Not exploded)' : 'Mesh search (exploded)';
if (/^["“”].*["“”]$/.test(leaf.content)) leaf.content = leaf.content.substr(1, leaf.content.length - 2); // Remove wrapping '"' characters
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^([^\s\/]*)\/?([^\s\)\/]+)?\[(majr|mesh major topic)(:NoExp|:no exp)?\]/i.exec(q)) {
// Major Mesh term - PubMed syntax (No Quotes)
var newLeaf = {};
newLeaf.type = 'mesh';
newLeaf.content = match[1];
if (match[2]) {
newLeaf.heading = match[2];
}
newLeaf.field = match[4] ? 'MeSH Major Topic search (Not exploded)' : 'MeSH Major Topic search (exploded)';
if (/^["“”].*["“”]$/.test(newLeaf.content)) newLeaf.content = newLeaf.content.substr(1, newLeaf.content.length - 2); // Remove wrapping '"' characters
branch.nodes.push(newLeaf);
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^"([^\\[]*)"\/?([^\s\)\/]+)?\[(majr|mesh major topic)(:NoExp|:no exp)?\]/i.exec(q)) {
// Major Mesh term - PubMed syntax (With Quotes)
var newLeaf = {};
newLeaf.type = 'mesh';
newLeaf.content = match[1];
if (match[2]) {
newLeaf.heading = match[2];
}
newLeaf.field = match[4] ? 'MeSH Major Topic search (Not exploded)' : 'MeSH Major Topic search (exploded)';
if (/^["“”].*["“”]$/.test(newLeaf.content)) newLeaf.content = newLeaf.content.substr(1, newLeaf.content.length - 2); // Remove wrapping '"' characters
branch.nodes.push(newLeaf);
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^\[(mesh subheading|sh)\]/i.exec(q)) {
// Mesh subheading search - PubMed syntax
leaf.type = 'mesh';
leaf.field = 'MeSH subheading search';
if (/^["“”].*["“”]$/.test(leaf.content)) leaf.content = leaf.content.substr(1, leaf.content.length - 2); // Remove wrapping '"' characters
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^\.(fs)\./i.exec(q)) {
// Mesh subheading search - Ovid syntax
var useLeaf;
if (_lodash["default"].isObject(leaf) && leaf.type == 'phrase') {
useLeaf = leaf;
} else if (_lodash["default"].isArray(leaf) && lastGroup) {
useLeaf = lastGroup;
}
leaf.type = 'mesh';
useLeaf.field = 'MeSH subheading search';
if (/^["“”].*["“”]$/.test(leaf.content)) leaf.content = leaf.content.substr(1, leaf.content.length - 2); // Remove wrapping '"' characters
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^\.(xm|sh)\./i.exec(q)) {
// Mesh search (field codes) - Ovid syntax
var useLeaf;
if (_lodash["default"].isObject(leaf) && leaf.type == 'phrase') {
useLeaf = leaf;
} else if (_lodash["default"].isArray(leaf) && lastGroup) {
useLeaf = lastGroup;
}
leaf.type = 'mesh';
useLeaf.field = match[1] === "xm" ? 'Mesh search (exploded)' : 'Mesh search (Not exploded)';
if (/^["“”].*["“”]$/.test(leaf.content)) leaf.content = leaf.content.substr(1, leaf.content.length - 2); // Remove wrapping '"' characters
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if ((match = /^(exp "([^*]*?)"\/)\s*/i.exec(q)) || (match = /^(exp ([^*]*?)\/([^\s\)]+)?)\s*/i.exec(q))) {
// Mesh term - Ovid syntax (exploded)
var newLeaf = {
type: 'mesh',
field: 'Mesh search (exploded)',
content: match[2]
};
if (match[3]) {
newLeaf.heading = match[3];
}
branch.nodes.push(newLeaf);
offset += match[1].length;
q = q.substr(match[1].length);
cropString = false;
afterWhitespace = true;
} else if ((match = /^(exp \*"([^*]*?)"\/)\s*/i.exec(q)) || (match = /^(exp \*([^*]*?)\/([^\s\)]+)?)\s*/i.exec(q))) {
// Major Mesh term - Ovid syntax
var newLeaf = {
type: 'mesh',
field: 'MeSH Major Topic search (exploded)',
content: match[2]
};
if (match[3]) {
newLeaf.heading = match[3];
}
branch.nodes.push(newLeaf);
offset += match[1].length;
q = q.substr(match[1].length);
cropString = false;
afterWhitespace = true;
} else if (/^\//.test(q) && leaf && leaf.type && leaf.type == 'phrase') {
// Mesh term - Ovid syntax (non-exploded)
leaf.type = 'mesh';
// Major Mesh
if (leaf.content[0] == "*") {
leaf.content = leaf.content.substr(1);
leaf.field = 'MeSH Major Topic search (Not exploded)';
} else leaf.field = 'Mesh search (Not exploded)';
}
// }}}
else if (match = /^<(.*?)>/.exec(q)) {
branch.nodes.push({
type: 'template',
content: match[1].toLowerCase()
});
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^(\n)+/.exec(q)) {
if (settings.preserveNewlines) {
var number_newline = match[0].length;
branch.nodes.push({
type: 'raw',
content: '\n'.repeat(number_newline)
});
leaf = undefined;
}
lineNumber += match[0].length;
if (branchStack.length > 0 && branchStack[branchStack.length - 1].nodes.every(function (node) {
return node.type !== "group";
})) {
// If we are currently inside a group don't add a newline group
newLine(lineNumber);
}
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
afterWhitespace = true;
} else if (match = /^(\r)+/.exec(q)) {
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
afterWhitespace = true;
}
// Match field codes {{{
else if (match = new RegExp("^(".concat(fieldCodes, ") *(\\[mp=[^\\]\\n]*\\])?"), "i").exec(q)) {
// Field specifier - PubMed syntax
// Figure out the leaf to use (usually the last one) or the previously used group
var useLeaf;
if (_lodash["default"].isObject(leaf) && leaf.type == 'phrase') {
useLeaf = leaf;
} else if (_lodash["default"].isArray(leaf) && lastGroup) {
useLeaf = lastGroup;
}
useLeaf.field = _fieldCodesParse["default"][match[1].toLowerCase()];
if (match[2]) {
offset += match[0].length;
q = q.substr(match[0].length);
} else {
offset += match[1].length;
q = q.substr(match[1].length);
}
cropString = false;
}
// }}}
/// Comment {{{
else if (match = /^#([^\)\d\n][^\)\n]+)/.exec(q)) {
trimLastLeaf();
branch.nodes.push({
type: 'comment',
content: match[1]
});
leaf = undefined;
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
}
// }}}
else {
var nextChar = q.substr(0, 1);
if ((_lodash["default"].isUndefined(leaf) || _lodash["default"].isArray(leaf)) && nextChar != ' ') {
// Leaf pointing to array entity - probably not created fallback leaf to append to
if (/^["“”]$/.test(nextChar) && (match = /^["“”](.*?)["“”]/.exec(q))) {
// First character is a speachmark - slurp until we see the next one
leaf = {
type: 'phrase',
content: match[1],
offset: offset
};
branch.nodes.push(leaf);
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else if (match = /^[^\s:/[.)]+/.exec(q)) {
// Slurp the phrase until the space or any character which indicates the end of a phrase
leaf = {
type: 'phrase',
content: match[0],
offset: offset
};
branch.nodes.push(leaf);
offset += match[0].length;
q = q.substr(match[0].length);
cropString = false;
} else {
// All other first chars - just dump into a buffer and let it fill slowly
leaf = {
type: 'phrase',
content: nextChar,
offset: offset
};
branch.nodes.push(leaf);
}
} else if (_lodash["default"].isObject(leaf) && leaf.type == 'phrase') {
leaf.content += nextChar;
}
afterWhitespace = nextChar == ' '; // Is the nextChar whitespace? Then set the flag
}
if (cropString) {
offset += 1; // Increment offset by 1
q = q.substr(1); // Crop 1 character
}
}
if (settings.transposeLines) {
_tools["default"].visit(tree.nodes, ['ref'], function (node, path) {
var reference;
var line;
// Find the matching line
for (reference in node.ref) {
var found = false;
for (line in tree.nodes) {
// If custom numbering is used only use nodes that are numbered by the user
if (userLineNumber) {
if (tree.nodes[line].number == node.ref[reference] && tree.nodes[line].isNumbered) {
// Copy the nodes from that line into the reference nodes
node.nodes.push(Array.from(tree.nodes[line].nodes));
// Pop the raw node
node.nodes[reference].pop();
found = true;
}
} else {
if (tree.nodes[line].number == node.ref[reference]) {
// Copy the nodes from that line into the reference nodes
node.nodes.push(Array.from(tree.nodes[line].nodes));
// Pop the raw node
node.nodes[reference].pop();
found = true;
}
}
}
// Line not found, push error message
if (!found) {
node.nodes.push([{
type: "phrase",
content: _tools["default"].createTooltip("Line " + node.ref[reference] + " not found", "Polyglot could not find specified line in the search query", "red-underline")
}]);
}
}
});
}
return tree.nodes;
};
exports.parse = parse;