clustal-js 2.0.0 → 2.0.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +4 -0
- package/README.md +2 -1
- package/dist/index.d.ts +35 -0
- package/dist/index.js +54 -0
- package/dist/index.js.map +1 -0
- package/dist/pairwise.d.ts +11 -0
- package/dist/pairwise.js +57 -0
- package/dist/pairwise.js.map +1 -0
- package/dist/util.d.ts +17 -0
- package/dist/util.js +78 -0
- package/dist/util.js.map +1 -0
- package/esm/index.d.ts +35 -0
- package/esm/index.js +48 -0
- package/esm/index.js.map +1 -0
- package/esm/pairwise.d.ts +11 -0
- package/esm/pairwise.js +52 -0
- package/esm/pairwise.js.map +1 -0
- package/esm/util.d.ts +17 -0
- package/esm/util.js +70 -0
- package/esm/util.js.map +1 -0
- package/package.json +21 -13
- package/src/index.ts +65 -0
- package/src/pairwise.ts +55 -0
- package/src/util.ts +79 -0
package/CHANGELOG.md
CHANGED
package/README.md
CHANGED
|
@@ -1,4 +1,5 @@
|
|
|
1
|
-
[](https://npmjs.org/package/@gmod/bam)
|
|
2
|
+
[](https://github.com/cmdcolin/clustal-js/actions?query=branch%3Amaster+workflow%3APush+)
|
|
2
3
|
|
|
3
4
|
# clustal-js
|
|
4
5
|
|
package/dist/index.d.ts
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
1
|
+
export declare function parseClustalIter(arr: Iterator<string>): {
|
|
2
|
+
consensus: string;
|
|
3
|
+
alns: {
|
|
4
|
+
id: string;
|
|
5
|
+
seq: string;
|
|
6
|
+
}[];
|
|
7
|
+
header: {
|
|
8
|
+
info: string;
|
|
9
|
+
version: string;
|
|
10
|
+
};
|
|
11
|
+
};
|
|
12
|
+
interface Row {
|
|
13
|
+
id: string;
|
|
14
|
+
seq: string;
|
|
15
|
+
}
|
|
16
|
+
export declare function parsePairwiseIter(arr: string): {
|
|
17
|
+
consensus: string;
|
|
18
|
+
alns: [Row, Row];
|
|
19
|
+
};
|
|
20
|
+
export declare function parse(contents: string): {
|
|
21
|
+
consensus: string;
|
|
22
|
+
alns: {
|
|
23
|
+
id: string;
|
|
24
|
+
seq: string;
|
|
25
|
+
}[];
|
|
26
|
+
header: {
|
|
27
|
+
info: string;
|
|
28
|
+
version: string;
|
|
29
|
+
};
|
|
30
|
+
};
|
|
31
|
+
export declare function parsePairwise(contents: string): {
|
|
32
|
+
consensus: string;
|
|
33
|
+
alns: [Row, Row];
|
|
34
|
+
};
|
|
35
|
+
export {};
|
package/dist/index.js
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.parseClustalIter = parseClustalIter;
|
|
4
|
+
exports.parsePairwiseIter = parsePairwiseIter;
|
|
5
|
+
exports.parse = parse;
|
|
6
|
+
exports.parsePairwise = parsePairwise;
|
|
7
|
+
const pairwise_1 = require("./pairwise");
|
|
8
|
+
const util_1 = require("./util");
|
|
9
|
+
function parseClustalIter(arr) {
|
|
10
|
+
const line = (0, util_1.getFirstNonEmptyLine)(arr);
|
|
11
|
+
if (!line) {
|
|
12
|
+
throw new Error('Empty file received');
|
|
13
|
+
}
|
|
14
|
+
const header = (0, util_1.parseHeader)(line);
|
|
15
|
+
const res = (0, util_1.parseBlocks)(arr);
|
|
16
|
+
if (res === undefined) {
|
|
17
|
+
throw new Error('No blocks parsed');
|
|
18
|
+
}
|
|
19
|
+
const alns = res.seqs.map((n, index) => ({ id: res.ids[index], seq: n }));
|
|
20
|
+
const { consensus } = res;
|
|
21
|
+
if (consensus.length != alns[0].seq.length) {
|
|
22
|
+
throw new Error(`Consensus length != sequence length. Con ${consensus.length} seq ${alns[0].seq.length}`);
|
|
23
|
+
}
|
|
24
|
+
return { consensus, alns, header };
|
|
25
|
+
}
|
|
26
|
+
function parsePairwiseIter(arr) {
|
|
27
|
+
const res = (0, pairwise_1.parsePairwiseBlocks)(arr.split('\n')[Symbol.iterator]());
|
|
28
|
+
if (res === undefined) {
|
|
29
|
+
throw new Error('No blocks parsed');
|
|
30
|
+
}
|
|
31
|
+
const alns = res.seqs.map((n, index) => ({ id: res.ids[index], seq: n }));
|
|
32
|
+
const { consensus } = res;
|
|
33
|
+
if (consensus.length != alns[0].seq.length) {
|
|
34
|
+
throw new Error(`Consensus length != sequence length. Con ${consensus.length} seq ${alns[0].seq.length}`);
|
|
35
|
+
}
|
|
36
|
+
if (alns.length !== 2) {
|
|
37
|
+
throw new Error('More than two sequences in pairwise alignment');
|
|
38
|
+
}
|
|
39
|
+
return {
|
|
40
|
+
consensus,
|
|
41
|
+
alns: alns,
|
|
42
|
+
};
|
|
43
|
+
}
|
|
44
|
+
function parse(contents) {
|
|
45
|
+
const iter = contents.split('\n')[Symbol.iterator]();
|
|
46
|
+
return parseClustalIter(iter);
|
|
47
|
+
}
|
|
48
|
+
function parsePairwise(contents) {
|
|
49
|
+
return parsePairwiseIter(contents
|
|
50
|
+
.split('\n')
|
|
51
|
+
.filter(f => !f.startsWith('#'))
|
|
52
|
+
.join('\n'));
|
|
53
|
+
}
|
|
54
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";;AAGA,4CAqBC;AAMD,8CAoBC;AAED,sBAGC;AAED,sCAOC;AAhED,yCAAgD;AAChD,iCAAuE;AAEvE,SAAgB,gBAAgB,CAAC,GAAqB;IACpD,MAAM,IAAI,GAAG,IAAA,2BAAoB,EAAC,GAAG,CAAC,CAAA;IACtC,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,MAAM,IAAI,KAAK,CAAC,qBAAqB,CAAC,CAAA;IACxC,CAAC;IACD,MAAM,MAAM,GAAG,IAAA,kBAAW,EAAC,IAAI,CAAC,CAAA;IAEhC,MAAM,GAAG,GAAG,IAAA,kBAAW,EAAC,GAAG,CAAC,CAAA;IAC5B,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,kBAAkB,CAAC,CAAA;IACrC,CAAC;IAED,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;IACzE,MAAM,EAAE,SAAS,EAAE,GAAG,GAAG,CAAA;IACzB,IAAI,SAAS,CAAC,MAAM,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;QAC3C,MAAM,IAAI,KAAK,CACb,4CAA4C,SAAS,CAAC,MAAM,QAAQ,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CACzF,CAAA;IACH,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;AACpC,CAAC;AAMD,SAAgB,iBAAiB,CAAC,GAAW;IAC3C,MAAM,GAAG,GAAG,IAAA,8BAAmB,EAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAA;IACnE,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,kBAAkB,CAAC,CAAA;IACrC,CAAC;IAED,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;IACzE,MAAM,EAAE,SAAS,EAAE,GAAG,GAAG,CAAA;IACzB,IAAI,SAAS,CAAC,MAAM,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;QAC3C,MAAM,IAAI,KAAK,CACb,4CAA4C,SAAS,CAAC,MAAM,QAAQ,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CACzF,CAAA;IACH,CAAC;IACD,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,+CAA+C,CAAC,CAAA;IAClE,CAAC;IACD,OAAO;QACL,SAAS;QACT,IAAI,EAAE,IAAkB;KACzB,CAAA;AACH,CAAC;AAED,SAAgB,KAAK,CAAC,QAAgB;IACpC,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAA;IACpD,OAAO,gBAAgB,CAAC,IAAI,CAAC,CAAA;AAC/B,CAAC;AAED,SAAgB,aAAa,CAAC,QAAgB;IAC5C,OAAO,iBAAiB,CACtB,QAAQ;SACL,KAAK,CAAC,IAAI,CAAC;SACX,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;SAC/B,IAAI,CAAC,IAAI,CAAC,CACd,CAAA;AACH,CAAC"}
|
|
@@ -0,0 +1,11 @@
|
|
|
1
|
+
export declare function getSeqBounds(line: string): readonly [number, number];
|
|
2
|
+
export declare function parsePairwiseBlock(arr: Iterator<string>): {
|
|
3
|
+
ids: string[];
|
|
4
|
+
seqs: string[];
|
|
5
|
+
consensus: string;
|
|
6
|
+
} | undefined;
|
|
7
|
+
export declare function parsePairwiseBlocks(arr: Iterator<string>): {
|
|
8
|
+
ids: string[];
|
|
9
|
+
seqs: string[];
|
|
10
|
+
consensus: string;
|
|
11
|
+
} | undefined;
|
package/dist/pairwise.js
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.getSeqBounds = getSeqBounds;
|
|
4
|
+
exports.parsePairwiseBlock = parsePairwiseBlock;
|
|
5
|
+
exports.parsePairwiseBlocks = parsePairwiseBlocks;
|
|
6
|
+
const util_1 = require("./util");
|
|
7
|
+
function getSeqBounds(line) {
|
|
8
|
+
const fields = line.split(/\s+/);
|
|
9
|
+
const k1 = fields[0].length + fields[1].length;
|
|
10
|
+
const temp = line.slice(k1);
|
|
11
|
+
const s = k1 + temp.indexOf(fields[2]);
|
|
12
|
+
const e = s + fields[2].length;
|
|
13
|
+
return [s, e];
|
|
14
|
+
}
|
|
15
|
+
// Use the first block to get the sequence identifiers
|
|
16
|
+
function parsePairwiseBlock(arr) {
|
|
17
|
+
let line = (0, util_1.getFirstNonEmptyLine)(arr);
|
|
18
|
+
const block = [];
|
|
19
|
+
let consensusLine = '';
|
|
20
|
+
if (!line) {
|
|
21
|
+
return undefined;
|
|
22
|
+
}
|
|
23
|
+
while (line) {
|
|
24
|
+
if (line.startsWith(' ')) {
|
|
25
|
+
consensusLine = line;
|
|
26
|
+
}
|
|
27
|
+
else {
|
|
28
|
+
block.push(line);
|
|
29
|
+
}
|
|
30
|
+
line = arr.next().value;
|
|
31
|
+
}
|
|
32
|
+
const [start, end] = getSeqBounds(block[0]);
|
|
33
|
+
const fields = block.map(s => s.split(/\s+/));
|
|
34
|
+
const ids = fields.map(s => s[0]);
|
|
35
|
+
const seqs = fields.map(s => s[2]);
|
|
36
|
+
let consensus = consensusLine.slice(start, end);
|
|
37
|
+
// handle if the consensus trailing whitespace got trimmed
|
|
38
|
+
const remainder = seqs[0].length - consensus.length;
|
|
39
|
+
if (remainder) {
|
|
40
|
+
consensus += ' '.repeat(remainder);
|
|
41
|
+
}
|
|
42
|
+
return { ids, seqs, consensus };
|
|
43
|
+
}
|
|
44
|
+
function parsePairwiseBlocks(arr) {
|
|
45
|
+
let block;
|
|
46
|
+
const res = parsePairwiseBlock(arr);
|
|
47
|
+
if (res !== undefined) {
|
|
48
|
+
while ((block = parsePairwiseBlock(arr))) {
|
|
49
|
+
for (let i = 0; i < block.seqs.length; i++) {
|
|
50
|
+
res.seqs[i] += block.seqs[i];
|
|
51
|
+
}
|
|
52
|
+
res.consensus += block.consensus;
|
|
53
|
+
}
|
|
54
|
+
}
|
|
55
|
+
return res;
|
|
56
|
+
}
|
|
57
|
+
//# sourceMappingURL=pairwise.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pairwise.js","sourceRoot":"","sources":["../src/pairwise.ts"],"names":[],"mappings":";;AAEA,oCAOC;AAGD,gDA4BC;AAED,kDAYC;AAtDD,iCAA6C;AAE7C,SAAgB,YAAY,CAAC,IAAY;IACvC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;IAChC,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;IAC3B,MAAM,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAA;IACtC,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC9B,OAAO,CAAC,CAAC,EAAE,CAAC,CAAU,CAAA;AACxB,CAAC;AAED,sDAAsD;AACtD,SAAgB,kBAAkB,CAAC,GAAqB;IACtD,IAAI,IAAI,GAAG,IAAA,2BAAoB,EAAC,GAAG,CAAC,CAAA;IACpC,MAAM,KAAK,GAAG,EAAE,CAAA;IAChB,IAAI,aAAa,GAAG,EAAE,CAAA;IACtB,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,OAAO,IAAI,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACzB,aAAa,GAAG,IAAI,CAAA;QACtB,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;QACD,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC,KAAK,CAAA;IACzB,CAAC;IACD,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;IAC3C,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAA;IAC7C,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IACjC,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAClC,IAAI,SAAS,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAA;IAE/C,0DAA0D;IAC1D,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,CAAA;IACnD,IAAI,SAAS,EAAE,CAAC;QACd,SAAS,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IACpC,CAAC;IACD,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAA;AACjC,CAAC;AAED,SAAgB,mBAAmB,CAAC,GAAqB;IACvD,IAAI,KAAK,CAAA;IACT,MAAM,GAAG,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAA;IACnC,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YACzC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YAC9B,CAAC;YACD,GAAG,CAAC,SAAS,IAAI,KAAK,CAAC,SAAS,CAAA;QAClC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}
|
package/dist/util.d.ts
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
export declare function parseVersion(line: string): string;
|
|
2
|
+
export declare function parseHeader(info: string): {
|
|
3
|
+
info: string;
|
|
4
|
+
version: string;
|
|
5
|
+
};
|
|
6
|
+
export declare function getFirstNonEmptyLine(arr: Iterator<string>): string;
|
|
7
|
+
export declare function getSeqBounds(line: string): readonly [number, number];
|
|
8
|
+
export declare function parseBlock(arr: Iterator<string>): {
|
|
9
|
+
ids: string[];
|
|
10
|
+
seqs: string[];
|
|
11
|
+
consensus: string;
|
|
12
|
+
} | undefined;
|
|
13
|
+
export declare function parseBlocks(arr: Iterator<string>): {
|
|
14
|
+
ids: string[];
|
|
15
|
+
seqs: string[];
|
|
16
|
+
consensus: string;
|
|
17
|
+
} | undefined;
|
package/dist/util.js
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.parseVersion = parseVersion;
|
|
4
|
+
exports.parseHeader = parseHeader;
|
|
5
|
+
exports.getFirstNonEmptyLine = getFirstNonEmptyLine;
|
|
6
|
+
exports.getSeqBounds = getSeqBounds;
|
|
7
|
+
exports.parseBlock = parseBlock;
|
|
8
|
+
exports.parseBlocks = parseBlocks;
|
|
9
|
+
function parseVersion(line) {
|
|
10
|
+
const res = /\(?(\d+(\.\d+)+)\)?/.exec(line);
|
|
11
|
+
return res && res.length > 1 ? res[1] : '';
|
|
12
|
+
}
|
|
13
|
+
function parseHeader(info) {
|
|
14
|
+
const knownHeaders = ['CLUSTAL', 'PROBCONS', 'MUSCLE', 'MSAPROBS', 'Kalign'];
|
|
15
|
+
if (!knownHeaders.some(l => info.startsWith(l))) {
|
|
16
|
+
console.warn(`${info} is not a known CLUSTAL header: ${knownHeaders.join(',')}, proceeding but could indicate an issue`);
|
|
17
|
+
}
|
|
18
|
+
const version = parseVersion(info);
|
|
19
|
+
return { info, version };
|
|
20
|
+
}
|
|
21
|
+
function getFirstNonEmptyLine(arr) {
|
|
22
|
+
// There should be two blank lines after the header line
|
|
23
|
+
let line = arr.next();
|
|
24
|
+
while (!line.done && line.value.trim() === '') {
|
|
25
|
+
line = arr.next();
|
|
26
|
+
}
|
|
27
|
+
return line.value;
|
|
28
|
+
}
|
|
29
|
+
function getSeqBounds(line) {
|
|
30
|
+
const fields = line.split(/\s+/);
|
|
31
|
+
const temp = line.slice(fields[0].length);
|
|
32
|
+
const s = fields[0].length + temp.indexOf(fields[1]);
|
|
33
|
+
const e = s + fields[1].length;
|
|
34
|
+
return [s, e];
|
|
35
|
+
}
|
|
36
|
+
// Use the first block to get the sequence identifiers
|
|
37
|
+
function parseBlock(arr) {
|
|
38
|
+
let line = getFirstNonEmptyLine(arr);
|
|
39
|
+
const block = [];
|
|
40
|
+
let consensusLine = '';
|
|
41
|
+
if (!line) {
|
|
42
|
+
return undefined;
|
|
43
|
+
}
|
|
44
|
+
while (line) {
|
|
45
|
+
if (line[0].startsWith(' ')) {
|
|
46
|
+
consensusLine = line;
|
|
47
|
+
}
|
|
48
|
+
else {
|
|
49
|
+
block.push(line);
|
|
50
|
+
}
|
|
51
|
+
line = arr.next().value;
|
|
52
|
+
}
|
|
53
|
+
const [start, end] = getSeqBounds(block[0]);
|
|
54
|
+
const fields = block.map(s => s.split(/\s+/));
|
|
55
|
+
const ids = fields.map(s => s[0]);
|
|
56
|
+
const seqs = block.map(s => s.slice(start, end));
|
|
57
|
+
let consensus = consensusLine.slice(start, end);
|
|
58
|
+
// handle if the consensus trailing whitespace got trimmed
|
|
59
|
+
const remainder = seqs[0].length - consensus.length;
|
|
60
|
+
if (remainder) {
|
|
61
|
+
consensus += ' '.repeat(remainder);
|
|
62
|
+
}
|
|
63
|
+
return { ids, seqs, consensus };
|
|
64
|
+
}
|
|
65
|
+
function parseBlocks(arr) {
|
|
66
|
+
let block;
|
|
67
|
+
const res = parseBlock(arr);
|
|
68
|
+
if (res !== undefined) {
|
|
69
|
+
while ((block = parseBlock(arr))) {
|
|
70
|
+
for (let i = 0; i < block.seqs.length; i++) {
|
|
71
|
+
res.seqs[i] += block.seqs[i];
|
|
72
|
+
}
|
|
73
|
+
res.consensus += block.consensus;
|
|
74
|
+
}
|
|
75
|
+
}
|
|
76
|
+
return res;
|
|
77
|
+
}
|
|
78
|
+
//# sourceMappingURL=util.js.map
|
package/dist/util.js.map
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"util.js","sourceRoot":"","sources":["../src/util.ts"],"names":[],"mappings":";;AAAA,oCAGC;AACD,kCAYC;AAED,oDAOC;AAED,oCAMC;AAGD,gCA4BC;AAED,kCAYC;AA9ED,SAAgB,YAAY,CAAC,IAAY;IACvC,MAAM,GAAG,GAAG,qBAAqB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAC5C,OAAO,GAAG,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;AAC5C,CAAC;AACD,SAAgB,WAAW,CAAC,IAAY;IACtC,MAAM,YAAY,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAA;IAE5E,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAChD,OAAO,CAAC,IAAI,CACV,GAAG,IAAI,mCAAmC,YAAY,CAAC,IAAI,CACzD,GAAG,CACJ,0CAA0C,CAC5C,CAAA;IACH,CAAC;IACD,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;IAClC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,CAAA;AAC1B,CAAC;AAED,SAAgB,oBAAoB,CAAC,GAAqB;IACxD,wDAAwD;IACxD,IAAI,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IACrB,OAAO,CAAC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;QAC9C,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IACnB,CAAC;IACD,OAAO,IAAI,CAAC,KAAK,CAAA;AACnB,CAAC;AAED,SAAgB,YAAY,CAAC,IAAY;IACvC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;IAChC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAA;IACpD,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC9B,OAAO,CAAC,CAAC,EAAE,CAAC,CAAU,CAAA;AACxB,CAAC;AAED,sDAAsD;AACtD,SAAgB,UAAU,CAAC,GAAqB;IAC9C,IAAI,IAAI,GAAG,oBAAoB,CAAC,GAAG,CAAC,CAAA;IACpC,MAAM,KAAK,GAAG,EAAE,CAAA;IAChB,IAAI,aAAa,GAAG,EAAE,CAAA;IACtB,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,OAAO,IAAI,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YAC5B,aAAa,GAAG,IAAI,CAAA;QACtB,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;QACD,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC,KAAK,CAAA;IACzB,CAAC;IACD,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;IAC3C,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAA;IAC7C,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IACjC,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAA;IAChD,IAAI,SAAS,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAA;IAE/C,0DAA0D;IAC1D,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,CAAA;IACnD,IAAI,SAAS,EAAE,CAAC;QACd,SAAS,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IACpC,CAAC;IACD,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAA;AACjC,CAAC;AAED,SAAgB,WAAW,CAAC,GAAqB;IAC/C,IAAI,KAAK,CAAA;IACT,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAA;IAC3B,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YAC9B,CAAC;YACD,GAAG,CAAC,SAAS,IAAI,KAAK,CAAC,SAAS,CAAA;QAClC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}
|
package/esm/index.d.ts
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
1
|
+
export declare function parseClustalIter(arr: Iterator<string>): {
|
|
2
|
+
consensus: string;
|
|
3
|
+
alns: {
|
|
4
|
+
id: string;
|
|
5
|
+
seq: string;
|
|
6
|
+
}[];
|
|
7
|
+
header: {
|
|
8
|
+
info: string;
|
|
9
|
+
version: string;
|
|
10
|
+
};
|
|
11
|
+
};
|
|
12
|
+
interface Row {
|
|
13
|
+
id: string;
|
|
14
|
+
seq: string;
|
|
15
|
+
}
|
|
16
|
+
export declare function parsePairwiseIter(arr: string): {
|
|
17
|
+
consensus: string;
|
|
18
|
+
alns: [Row, Row];
|
|
19
|
+
};
|
|
20
|
+
export declare function parse(contents: string): {
|
|
21
|
+
consensus: string;
|
|
22
|
+
alns: {
|
|
23
|
+
id: string;
|
|
24
|
+
seq: string;
|
|
25
|
+
}[];
|
|
26
|
+
header: {
|
|
27
|
+
info: string;
|
|
28
|
+
version: string;
|
|
29
|
+
};
|
|
30
|
+
};
|
|
31
|
+
export declare function parsePairwise(contents: string): {
|
|
32
|
+
consensus: string;
|
|
33
|
+
alns: [Row, Row];
|
|
34
|
+
};
|
|
35
|
+
export {};
|
package/esm/index.js
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
1
|
+
import { parsePairwiseBlocks } from './pairwise';
|
|
2
|
+
import { parseBlocks, parseHeader, getFirstNonEmptyLine } from './util';
|
|
3
|
+
export function parseClustalIter(arr) {
|
|
4
|
+
const line = getFirstNonEmptyLine(arr);
|
|
5
|
+
if (!line) {
|
|
6
|
+
throw new Error('Empty file received');
|
|
7
|
+
}
|
|
8
|
+
const header = parseHeader(line);
|
|
9
|
+
const res = parseBlocks(arr);
|
|
10
|
+
if (res === undefined) {
|
|
11
|
+
throw new Error('No blocks parsed');
|
|
12
|
+
}
|
|
13
|
+
const alns = res.seqs.map((n, index) => ({ id: res.ids[index], seq: n }));
|
|
14
|
+
const { consensus } = res;
|
|
15
|
+
if (consensus.length != alns[0].seq.length) {
|
|
16
|
+
throw new Error(`Consensus length != sequence length. Con ${consensus.length} seq ${alns[0].seq.length}`);
|
|
17
|
+
}
|
|
18
|
+
return { consensus, alns, header };
|
|
19
|
+
}
|
|
20
|
+
export function parsePairwiseIter(arr) {
|
|
21
|
+
const res = parsePairwiseBlocks(arr.split('\n')[Symbol.iterator]());
|
|
22
|
+
if (res === undefined) {
|
|
23
|
+
throw new Error('No blocks parsed');
|
|
24
|
+
}
|
|
25
|
+
const alns = res.seqs.map((n, index) => ({ id: res.ids[index], seq: n }));
|
|
26
|
+
const { consensus } = res;
|
|
27
|
+
if (consensus.length != alns[0].seq.length) {
|
|
28
|
+
throw new Error(`Consensus length != sequence length. Con ${consensus.length} seq ${alns[0].seq.length}`);
|
|
29
|
+
}
|
|
30
|
+
if (alns.length !== 2) {
|
|
31
|
+
throw new Error('More than two sequences in pairwise alignment');
|
|
32
|
+
}
|
|
33
|
+
return {
|
|
34
|
+
consensus,
|
|
35
|
+
alns: alns,
|
|
36
|
+
};
|
|
37
|
+
}
|
|
38
|
+
export function parse(contents) {
|
|
39
|
+
const iter = contents.split('\n')[Symbol.iterator]();
|
|
40
|
+
return parseClustalIter(iter);
|
|
41
|
+
}
|
|
42
|
+
export function parsePairwise(contents) {
|
|
43
|
+
return parsePairwiseIter(contents
|
|
44
|
+
.split('\n')
|
|
45
|
+
.filter(f => !f.startsWith('#'))
|
|
46
|
+
.join('\n'));
|
|
47
|
+
}
|
|
48
|
+
//# sourceMappingURL=index.js.map
|
package/esm/index.js.map
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,YAAY,CAAA;AAChD,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,oBAAoB,EAAE,MAAM,QAAQ,CAAA;AAEvE,MAAM,UAAU,gBAAgB,CAAC,GAAqB;IACpD,MAAM,IAAI,GAAG,oBAAoB,CAAC,GAAG,CAAC,CAAA;IACtC,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,MAAM,IAAI,KAAK,CAAC,qBAAqB,CAAC,CAAA;IACxC,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,IAAI,CAAC,CAAA;IAEhC,MAAM,GAAG,GAAG,WAAW,CAAC,GAAG,CAAC,CAAA;IAC5B,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,kBAAkB,CAAC,CAAA;IACrC,CAAC;IAED,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;IACzE,MAAM,EAAE,SAAS,EAAE,GAAG,GAAG,CAAA;IACzB,IAAI,SAAS,CAAC,MAAM,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;QAC3C,MAAM,IAAI,KAAK,CACb,4CAA4C,SAAS,CAAC,MAAM,QAAQ,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CACzF,CAAA;IACH,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;AACpC,CAAC;AAMD,MAAM,UAAU,iBAAiB,CAAC,GAAW;IAC3C,MAAM,GAAG,GAAG,mBAAmB,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAA;IACnE,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,kBAAkB,CAAC,CAAA;IACrC,CAAC;IAED,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;IACzE,MAAM,EAAE,SAAS,EAAE,GAAG,GAAG,CAAA;IACzB,IAAI,SAAS,CAAC,MAAM,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;QAC3C,MAAM,IAAI,KAAK,CACb,4CAA4C,SAAS,CAAC,MAAM,QAAQ,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,EAAE,CACzF,CAAA;IACH,CAAC;IACD,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACtB,MAAM,IAAI,KAAK,CAAC,+CAA+C,CAAC,CAAA;IAClE,CAAC;IACD,OAAO;QACL,SAAS;QACT,IAAI,EAAE,IAAkB;KACzB,CAAA;AACH,CAAC;AAED,MAAM,UAAU,KAAK,CAAC,QAAgB;IACpC,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAA;IACpD,OAAO,gBAAgB,CAAC,IAAI,CAAC,CAAA;AAC/B,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,QAAgB;IAC5C,OAAO,iBAAiB,CACtB,QAAQ;SACL,KAAK,CAAC,IAAI,CAAC;SACX,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;SAC/B,IAAI,CAAC,IAAI,CAAC,CACd,CAAA;AACH,CAAC"}
|
|
@@ -0,0 +1,11 @@
|
|
|
1
|
+
export declare function getSeqBounds(line: string): readonly [number, number];
|
|
2
|
+
export declare function parsePairwiseBlock(arr: Iterator<string>): {
|
|
3
|
+
ids: string[];
|
|
4
|
+
seqs: string[];
|
|
5
|
+
consensus: string;
|
|
6
|
+
} | undefined;
|
|
7
|
+
export declare function parsePairwiseBlocks(arr: Iterator<string>): {
|
|
8
|
+
ids: string[];
|
|
9
|
+
seqs: string[];
|
|
10
|
+
consensus: string;
|
|
11
|
+
} | undefined;
|
package/esm/pairwise.js
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
import { getFirstNonEmptyLine } from './util';
|
|
2
|
+
export function getSeqBounds(line) {
|
|
3
|
+
const fields = line.split(/\s+/);
|
|
4
|
+
const k1 = fields[0].length + fields[1].length;
|
|
5
|
+
const temp = line.slice(k1);
|
|
6
|
+
const s = k1 + temp.indexOf(fields[2]);
|
|
7
|
+
const e = s + fields[2].length;
|
|
8
|
+
return [s, e];
|
|
9
|
+
}
|
|
10
|
+
// Use the first block to get the sequence identifiers
|
|
11
|
+
export function parsePairwiseBlock(arr) {
|
|
12
|
+
let line = getFirstNonEmptyLine(arr);
|
|
13
|
+
const block = [];
|
|
14
|
+
let consensusLine = '';
|
|
15
|
+
if (!line) {
|
|
16
|
+
return undefined;
|
|
17
|
+
}
|
|
18
|
+
while (line) {
|
|
19
|
+
if (line.startsWith(' ')) {
|
|
20
|
+
consensusLine = line;
|
|
21
|
+
}
|
|
22
|
+
else {
|
|
23
|
+
block.push(line);
|
|
24
|
+
}
|
|
25
|
+
line = arr.next().value;
|
|
26
|
+
}
|
|
27
|
+
const [start, end] = getSeqBounds(block[0]);
|
|
28
|
+
const fields = block.map(s => s.split(/\s+/));
|
|
29
|
+
const ids = fields.map(s => s[0]);
|
|
30
|
+
const seqs = fields.map(s => s[2]);
|
|
31
|
+
let consensus = consensusLine.slice(start, end);
|
|
32
|
+
// handle if the consensus trailing whitespace got trimmed
|
|
33
|
+
const remainder = seqs[0].length - consensus.length;
|
|
34
|
+
if (remainder) {
|
|
35
|
+
consensus += ' '.repeat(remainder);
|
|
36
|
+
}
|
|
37
|
+
return { ids, seqs, consensus };
|
|
38
|
+
}
|
|
39
|
+
export function parsePairwiseBlocks(arr) {
|
|
40
|
+
let block;
|
|
41
|
+
const res = parsePairwiseBlock(arr);
|
|
42
|
+
if (res !== undefined) {
|
|
43
|
+
while ((block = parsePairwiseBlock(arr))) {
|
|
44
|
+
for (let i = 0; i < block.seqs.length; i++) {
|
|
45
|
+
res.seqs[i] += block.seqs[i];
|
|
46
|
+
}
|
|
47
|
+
res.consensus += block.consensus;
|
|
48
|
+
}
|
|
49
|
+
}
|
|
50
|
+
return res;
|
|
51
|
+
}
|
|
52
|
+
//# sourceMappingURL=pairwise.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pairwise.js","sourceRoot":"","sources":["../src/pairwise.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,QAAQ,CAAA;AAE7C,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;IAChC,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;IAC3B,MAAM,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAA;IACtC,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC9B,OAAO,CAAC,CAAC,EAAE,CAAC,CAAU,CAAA;AACxB,CAAC;AAED,sDAAsD;AACtD,MAAM,UAAU,kBAAkB,CAAC,GAAqB;IACtD,IAAI,IAAI,GAAG,oBAAoB,CAAC,GAAG,CAAC,CAAA;IACpC,MAAM,KAAK,GAAG,EAAE,CAAA;IAChB,IAAI,aAAa,GAAG,EAAE,CAAA;IACtB,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,OAAO,IAAI,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACzB,aAAa,GAAG,IAAI,CAAA;QACtB,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;QACD,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC,KAAK,CAAA;IACzB,CAAC;IACD,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;IAC3C,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAA;IAC7C,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IACjC,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAClC,IAAI,SAAS,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAA;IAE/C,0DAA0D;IAC1D,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,CAAA;IACnD,IAAI,SAAS,EAAE,CAAC;QACd,SAAS,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IACpC,CAAC;IACD,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAA;AACjC,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,GAAqB;IACvD,IAAI,KAAK,CAAA;IACT,MAAM,GAAG,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAA;IACnC,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YACzC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YAC9B,CAAC;YACD,GAAG,CAAC,SAAS,IAAI,KAAK,CAAC,SAAS,CAAA;QAClC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}
|
package/esm/util.d.ts
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
export declare function parseVersion(line: string): string;
|
|
2
|
+
export declare function parseHeader(info: string): {
|
|
3
|
+
info: string;
|
|
4
|
+
version: string;
|
|
5
|
+
};
|
|
6
|
+
export declare function getFirstNonEmptyLine(arr: Iterator<string>): string;
|
|
7
|
+
export declare function getSeqBounds(line: string): readonly [number, number];
|
|
8
|
+
export declare function parseBlock(arr: Iterator<string>): {
|
|
9
|
+
ids: string[];
|
|
10
|
+
seqs: string[];
|
|
11
|
+
consensus: string;
|
|
12
|
+
} | undefined;
|
|
13
|
+
export declare function parseBlocks(arr: Iterator<string>): {
|
|
14
|
+
ids: string[];
|
|
15
|
+
seqs: string[];
|
|
16
|
+
consensus: string;
|
|
17
|
+
} | undefined;
|
package/esm/util.js
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
1
|
+
export function parseVersion(line) {
|
|
2
|
+
const res = /\(?(\d+(\.\d+)+)\)?/.exec(line);
|
|
3
|
+
return res && res.length > 1 ? res[1] : '';
|
|
4
|
+
}
|
|
5
|
+
export function parseHeader(info) {
|
|
6
|
+
const knownHeaders = ['CLUSTAL', 'PROBCONS', 'MUSCLE', 'MSAPROBS', 'Kalign'];
|
|
7
|
+
if (!knownHeaders.some(l => info.startsWith(l))) {
|
|
8
|
+
console.warn(`${info} is not a known CLUSTAL header: ${knownHeaders.join(',')}, proceeding but could indicate an issue`);
|
|
9
|
+
}
|
|
10
|
+
const version = parseVersion(info);
|
|
11
|
+
return { info, version };
|
|
12
|
+
}
|
|
13
|
+
export function getFirstNonEmptyLine(arr) {
|
|
14
|
+
// There should be two blank lines after the header line
|
|
15
|
+
let line = arr.next();
|
|
16
|
+
while (!line.done && line.value.trim() === '') {
|
|
17
|
+
line = arr.next();
|
|
18
|
+
}
|
|
19
|
+
return line.value;
|
|
20
|
+
}
|
|
21
|
+
export function getSeqBounds(line) {
|
|
22
|
+
const fields = line.split(/\s+/);
|
|
23
|
+
const temp = line.slice(fields[0].length);
|
|
24
|
+
const s = fields[0].length + temp.indexOf(fields[1]);
|
|
25
|
+
const e = s + fields[1].length;
|
|
26
|
+
return [s, e];
|
|
27
|
+
}
|
|
28
|
+
// Use the first block to get the sequence identifiers
|
|
29
|
+
export function parseBlock(arr) {
|
|
30
|
+
let line = getFirstNonEmptyLine(arr);
|
|
31
|
+
const block = [];
|
|
32
|
+
let consensusLine = '';
|
|
33
|
+
if (!line) {
|
|
34
|
+
return undefined;
|
|
35
|
+
}
|
|
36
|
+
while (line) {
|
|
37
|
+
if (line[0].startsWith(' ')) {
|
|
38
|
+
consensusLine = line;
|
|
39
|
+
}
|
|
40
|
+
else {
|
|
41
|
+
block.push(line);
|
|
42
|
+
}
|
|
43
|
+
line = arr.next().value;
|
|
44
|
+
}
|
|
45
|
+
const [start, end] = getSeqBounds(block[0]);
|
|
46
|
+
const fields = block.map(s => s.split(/\s+/));
|
|
47
|
+
const ids = fields.map(s => s[0]);
|
|
48
|
+
const seqs = block.map(s => s.slice(start, end));
|
|
49
|
+
let consensus = consensusLine.slice(start, end);
|
|
50
|
+
// handle if the consensus trailing whitespace got trimmed
|
|
51
|
+
const remainder = seqs[0].length - consensus.length;
|
|
52
|
+
if (remainder) {
|
|
53
|
+
consensus += ' '.repeat(remainder);
|
|
54
|
+
}
|
|
55
|
+
return { ids, seqs, consensus };
|
|
56
|
+
}
|
|
57
|
+
export function parseBlocks(arr) {
|
|
58
|
+
let block;
|
|
59
|
+
const res = parseBlock(arr);
|
|
60
|
+
if (res !== undefined) {
|
|
61
|
+
while ((block = parseBlock(arr))) {
|
|
62
|
+
for (let i = 0; i < block.seqs.length; i++) {
|
|
63
|
+
res.seqs[i] += block.seqs[i];
|
|
64
|
+
}
|
|
65
|
+
res.consensus += block.consensus;
|
|
66
|
+
}
|
|
67
|
+
}
|
|
68
|
+
return res;
|
|
69
|
+
}
|
|
70
|
+
//# sourceMappingURL=util.js.map
|
package/esm/util.js.map
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"util.js","sourceRoot":"","sources":["../src/util.ts"],"names":[],"mappings":"AAAA,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,GAAG,GAAG,qBAAqB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IAC5C,OAAO,GAAG,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;AAC5C,CAAC;AACD,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,MAAM,YAAY,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAA;IAE5E,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAChD,OAAO,CAAC,IAAI,CACV,GAAG,IAAI,mCAAmC,YAAY,CAAC,IAAI,CACzD,GAAG,CACJ,0CAA0C,CAC5C,CAAA;IACH,CAAC;IACD,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;IAClC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,CAAA;AAC1B,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAC,GAAqB;IACxD,wDAAwD;IACxD,IAAI,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IACrB,OAAO,CAAC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;QAC9C,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IACnB,CAAC;IACD,OAAO,IAAI,CAAC,KAAK,CAAA;AACnB,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;IAChC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAA;IACpD,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;IAC9B,OAAO,CAAC,CAAC,EAAE,CAAC,CAAU,CAAA;AACxB,CAAC;AAED,sDAAsD;AACtD,MAAM,UAAU,UAAU,CAAC,GAAqB;IAC9C,IAAI,IAAI,GAAG,oBAAoB,CAAC,GAAG,CAAC,CAAA;IACpC,MAAM,KAAK,GAAG,EAAE,CAAA;IAChB,IAAI,aAAa,GAAG,EAAE,CAAA;IACtB,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,OAAO,IAAI,EAAE,CAAC;QACZ,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YAC5B,aAAa,GAAG,IAAI,CAAA;QACtB,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;QACD,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC,KAAK,CAAA;IACzB,CAAC;IACD,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;IAC3C,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAA;IAC7C,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IACjC,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAA;IAChD,IAAI,SAAS,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAA;IAE/C,0DAA0D;IAC1D,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,CAAA;IACnD,IAAI,SAAS,EAAE,CAAC;QACd,SAAS,IAAI,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IACpC,CAAC;IACD,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAA;AACjC,CAAC;AAED,MAAM,UAAU,WAAW,CAAC,GAAqB;IAC/C,IAAI,KAAK,CAAA;IACT,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAA;IAC3B,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YAC9B,CAAC;YACD,GAAG,CAAC,SAAS,IAAI,KAAK,CAAC,SAAS,CAAA;QAClC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}
|
package/package.json
CHANGED
|
@@ -1,30 +1,38 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "clustal-js",
|
|
3
|
-
"version": "2.0.
|
|
3
|
+
"version": "2.0.2",
|
|
4
4
|
"main": "dist/index.js",
|
|
5
|
+
"module": "esm/index.js",
|
|
5
6
|
"repository": "cmdcolin/clustal-js",
|
|
6
7
|
"files": [
|
|
7
|
-
"dist"
|
|
8
|
+
"dist",
|
|
9
|
+
"esm",
|
|
10
|
+
"src"
|
|
8
11
|
],
|
|
9
12
|
"license": "MIT",
|
|
10
13
|
"devDependencies": {
|
|
11
14
|
"@types/jest": "^29.5.12",
|
|
12
|
-
"@types/node": "^
|
|
13
|
-
"@typescript-eslint/eslint-plugin": "^
|
|
14
|
-
"@typescript-eslint/parser": "^
|
|
15
|
-
"eslint": "^
|
|
16
|
-
"eslint-
|
|
17
|
-
"eslint-plugin-prettier": "^5.1.3",
|
|
18
|
-
"eslint-plugin-unicorn": "^50.0.1",
|
|
15
|
+
"@types/node": "^22.0.0",
|
|
16
|
+
"@typescript-eslint/eslint-plugin": "^8.1.0",
|
|
17
|
+
"@typescript-eslint/parser": "^8.1.0",
|
|
18
|
+
"eslint": "^9.9.0",
|
|
19
|
+
"eslint-plugin-unicorn": "^55.0.0",
|
|
19
20
|
"jest": "^29.7.0",
|
|
20
21
|
"prettier": "^3.2.5",
|
|
21
|
-
"
|
|
22
|
-
"
|
|
22
|
+
"rimraf": "^6.0.1",
|
|
23
|
+
"ts-jest": "^29.2.4",
|
|
24
|
+
"typescript": "^5.3.3",
|
|
25
|
+
"typescript-eslint": "^8.1.0"
|
|
23
26
|
},
|
|
24
27
|
"scripts": {
|
|
28
|
+
"clean": "rimraf dist esm",
|
|
25
29
|
"test": "jest",
|
|
26
|
-
"lint": "eslint --
|
|
27
|
-
"build": "
|
|
30
|
+
"lint": "eslint --report-unused-disable-directives --max-warnings 0",
|
|
31
|
+
"build": "npm run build:esm && npm run build:es5",
|
|
32
|
+
"build:esm": "tsc --target es2018 --outDir esm",
|
|
33
|
+
"build:es5": "tsc --target es2015 --module commonjs --outDir dist",
|
|
34
|
+
"prebuild": "npm run clean",
|
|
35
|
+
"preversion": "npm run lint && npm test && npm run build",
|
|
28
36
|
"postversion": "git push --follow-tags"
|
|
29
37
|
}
|
|
30
38
|
}
|
package/src/index.ts
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
1
|
+
import { parsePairwiseBlocks } from './pairwise'
|
|
2
|
+
import { parseBlocks, parseHeader, getFirstNonEmptyLine } from './util'
|
|
3
|
+
|
|
4
|
+
export function parseClustalIter(arr: Iterator<string>) {
|
|
5
|
+
const line = getFirstNonEmptyLine(arr)
|
|
6
|
+
if (!line) {
|
|
7
|
+
throw new Error('Empty file received')
|
|
8
|
+
}
|
|
9
|
+
const header = parseHeader(line)
|
|
10
|
+
|
|
11
|
+
const res = parseBlocks(arr)
|
|
12
|
+
if (res === undefined) {
|
|
13
|
+
throw new Error('No blocks parsed')
|
|
14
|
+
}
|
|
15
|
+
|
|
16
|
+
const alns = res.seqs.map((n, index) => ({ id: res.ids[index], seq: n }))
|
|
17
|
+
const { consensus } = res
|
|
18
|
+
if (consensus.length != alns[0].seq.length) {
|
|
19
|
+
throw new Error(
|
|
20
|
+
`Consensus length != sequence length. Con ${consensus.length} seq ${alns[0].seq.length}`,
|
|
21
|
+
)
|
|
22
|
+
}
|
|
23
|
+
|
|
24
|
+
return { consensus, alns, header }
|
|
25
|
+
}
|
|
26
|
+
|
|
27
|
+
interface Row {
|
|
28
|
+
id: string
|
|
29
|
+
seq: string
|
|
30
|
+
}
|
|
31
|
+
export function parsePairwiseIter(arr: string) {
|
|
32
|
+
const res = parsePairwiseBlocks(arr.split('\n')[Symbol.iterator]())
|
|
33
|
+
if (res === undefined) {
|
|
34
|
+
throw new Error('No blocks parsed')
|
|
35
|
+
}
|
|
36
|
+
|
|
37
|
+
const alns = res.seqs.map((n, index) => ({ id: res.ids[index], seq: n }))
|
|
38
|
+
const { consensus } = res
|
|
39
|
+
if (consensus.length != alns[0].seq.length) {
|
|
40
|
+
throw new Error(
|
|
41
|
+
`Consensus length != sequence length. Con ${consensus.length} seq ${alns[0].seq.length}`,
|
|
42
|
+
)
|
|
43
|
+
}
|
|
44
|
+
if (alns.length !== 2) {
|
|
45
|
+
throw new Error('More than two sequences in pairwise alignment')
|
|
46
|
+
}
|
|
47
|
+
return {
|
|
48
|
+
consensus,
|
|
49
|
+
alns: alns as [Row, Row],
|
|
50
|
+
}
|
|
51
|
+
}
|
|
52
|
+
|
|
53
|
+
export function parse(contents: string) {
|
|
54
|
+
const iter = contents.split('\n')[Symbol.iterator]()
|
|
55
|
+
return parseClustalIter(iter)
|
|
56
|
+
}
|
|
57
|
+
|
|
58
|
+
export function parsePairwise(contents: string) {
|
|
59
|
+
return parsePairwiseIter(
|
|
60
|
+
contents
|
|
61
|
+
.split('\n')
|
|
62
|
+
.filter(f => !f.startsWith('#'))
|
|
63
|
+
.join('\n'),
|
|
64
|
+
)
|
|
65
|
+
}
|
package/src/pairwise.ts
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
1
|
+
import { getFirstNonEmptyLine } from './util'
|
|
2
|
+
|
|
3
|
+
export function getSeqBounds(line: string) {
|
|
4
|
+
const fields = line.split(/\s+/)
|
|
5
|
+
const k1 = fields[0].length + fields[1].length
|
|
6
|
+
const temp = line.slice(k1)
|
|
7
|
+
const s = k1 + temp.indexOf(fields[2])
|
|
8
|
+
const e = s + fields[2].length
|
|
9
|
+
return [s, e] as const
|
|
10
|
+
}
|
|
11
|
+
|
|
12
|
+
// Use the first block to get the sequence identifiers
|
|
13
|
+
export function parsePairwiseBlock(arr: Iterator<string>) {
|
|
14
|
+
let line = getFirstNonEmptyLine(arr)
|
|
15
|
+
const block = []
|
|
16
|
+
let consensusLine = ''
|
|
17
|
+
if (!line) {
|
|
18
|
+
return undefined
|
|
19
|
+
}
|
|
20
|
+
|
|
21
|
+
while (line) {
|
|
22
|
+
if (line.startsWith(' ')) {
|
|
23
|
+
consensusLine = line
|
|
24
|
+
} else {
|
|
25
|
+
block.push(line)
|
|
26
|
+
}
|
|
27
|
+
line = arr.next().value
|
|
28
|
+
}
|
|
29
|
+
const [start, end] = getSeqBounds(block[0])
|
|
30
|
+
const fields = block.map(s => s.split(/\s+/))
|
|
31
|
+
const ids = fields.map(s => s[0])
|
|
32
|
+
const seqs = fields.map(s => s[2])
|
|
33
|
+
let consensus = consensusLine.slice(start, end)
|
|
34
|
+
|
|
35
|
+
// handle if the consensus trailing whitespace got trimmed
|
|
36
|
+
const remainder = seqs[0].length - consensus.length
|
|
37
|
+
if (remainder) {
|
|
38
|
+
consensus += ' '.repeat(remainder)
|
|
39
|
+
}
|
|
40
|
+
return { ids, seqs, consensus }
|
|
41
|
+
}
|
|
42
|
+
|
|
43
|
+
export function parsePairwiseBlocks(arr: Iterator<string>) {
|
|
44
|
+
let block
|
|
45
|
+
const res = parsePairwiseBlock(arr)
|
|
46
|
+
if (res !== undefined) {
|
|
47
|
+
while ((block = parsePairwiseBlock(arr))) {
|
|
48
|
+
for (let i = 0; i < block.seqs.length; i++) {
|
|
49
|
+
res.seqs[i] += block.seqs[i]
|
|
50
|
+
}
|
|
51
|
+
res.consensus += block.consensus
|
|
52
|
+
}
|
|
53
|
+
}
|
|
54
|
+
return res
|
|
55
|
+
}
|
package/src/util.ts
ADDED
|
@@ -0,0 +1,79 @@
|
|
|
1
|
+
export function parseVersion(line: string) {
|
|
2
|
+
const res = /\(?(\d+(\.\d+)+)\)?/.exec(line)
|
|
3
|
+
return res && res.length > 1 ? res[1] : ''
|
|
4
|
+
}
|
|
5
|
+
export function parseHeader(info: string) {
|
|
6
|
+
const knownHeaders = ['CLUSTAL', 'PROBCONS', 'MUSCLE', 'MSAPROBS', 'Kalign']
|
|
7
|
+
|
|
8
|
+
if (!knownHeaders.some(l => info.startsWith(l))) {
|
|
9
|
+
console.warn(
|
|
10
|
+
`${info} is not a known CLUSTAL header: ${knownHeaders.join(
|
|
11
|
+
',',
|
|
12
|
+
)}, proceeding but could indicate an issue`,
|
|
13
|
+
)
|
|
14
|
+
}
|
|
15
|
+
const version = parseVersion(info)
|
|
16
|
+
return { info, version }
|
|
17
|
+
}
|
|
18
|
+
|
|
19
|
+
export function getFirstNonEmptyLine(arr: Iterator<string>): string {
|
|
20
|
+
// There should be two blank lines after the header line
|
|
21
|
+
let line = arr.next()
|
|
22
|
+
while (!line.done && line.value.trim() === '') {
|
|
23
|
+
line = arr.next()
|
|
24
|
+
}
|
|
25
|
+
return line.value
|
|
26
|
+
}
|
|
27
|
+
|
|
28
|
+
export function getSeqBounds(line: string) {
|
|
29
|
+
const fields = line.split(/\s+/)
|
|
30
|
+
const temp = line.slice(fields[0].length)
|
|
31
|
+
const s = fields[0].length + temp.indexOf(fields[1])
|
|
32
|
+
const e = s + fields[1].length
|
|
33
|
+
return [s, e] as const
|
|
34
|
+
}
|
|
35
|
+
|
|
36
|
+
// Use the first block to get the sequence identifiers
|
|
37
|
+
export function parseBlock(arr: Iterator<string>) {
|
|
38
|
+
let line = getFirstNonEmptyLine(arr)
|
|
39
|
+
const block = []
|
|
40
|
+
let consensusLine = ''
|
|
41
|
+
if (!line) {
|
|
42
|
+
return undefined
|
|
43
|
+
}
|
|
44
|
+
|
|
45
|
+
while (line) {
|
|
46
|
+
if (line[0].startsWith(' ')) {
|
|
47
|
+
consensusLine = line
|
|
48
|
+
} else {
|
|
49
|
+
block.push(line)
|
|
50
|
+
}
|
|
51
|
+
line = arr.next().value
|
|
52
|
+
}
|
|
53
|
+
const [start, end] = getSeqBounds(block[0])
|
|
54
|
+
const fields = block.map(s => s.split(/\s+/))
|
|
55
|
+
const ids = fields.map(s => s[0])
|
|
56
|
+
const seqs = block.map(s => s.slice(start, end))
|
|
57
|
+
let consensus = consensusLine.slice(start, end)
|
|
58
|
+
|
|
59
|
+
// handle if the consensus trailing whitespace got trimmed
|
|
60
|
+
const remainder = seqs[0].length - consensus.length
|
|
61
|
+
if (remainder) {
|
|
62
|
+
consensus += ' '.repeat(remainder)
|
|
63
|
+
}
|
|
64
|
+
return { ids, seqs, consensus }
|
|
65
|
+
}
|
|
66
|
+
|
|
67
|
+
export function parseBlocks(arr: Iterator<string>) {
|
|
68
|
+
let block
|
|
69
|
+
const res = parseBlock(arr)
|
|
70
|
+
if (res !== undefined) {
|
|
71
|
+
while ((block = parseBlock(arr))) {
|
|
72
|
+
for (let i = 0; i < block.seqs.length; i++) {
|
|
73
|
+
res.seqs[i] += block.seqs[i]
|
|
74
|
+
}
|
|
75
|
+
res.consensus += block.consensus
|
|
76
|
+
}
|
|
77
|
+
}
|
|
78
|
+
return res
|
|
79
|
+
}
|