msa-parsers 1.0.3

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.
Files changed (85) hide show
  1. package/dist/gff/gffToInterPro.d.ts +20 -0
  2. package/dist/gff/gffToInterPro.js +83 -0
  3. package/dist/gff/gffToInterPro.js.map +1 -0
  4. package/dist/gff/gffToInterPro.test.d.ts +1 -0
  5. package/dist/gff/gffToInterPro.test.js +181 -0
  6. package/dist/gff/gffToInterPro.test.js.map +1 -0
  7. package/dist/gff/index.d.ts +3 -0
  8. package/dist/gff/index.js +4 -0
  9. package/dist/gff/index.js.map +1 -0
  10. package/dist/gff/interProToGFF.d.ts +9 -0
  11. package/dist/gff/interProToGFF.js +48 -0
  12. package/dist/gff/interProToGFF.js.map +1 -0
  13. package/dist/gff/interProToGFF.test.d.ts +1 -0
  14. package/dist/gff/interProToGFF.test.js +189 -0
  15. package/dist/gff/interProToGFF.test.js.map +1 -0
  16. package/dist/gff/parseGFF.d.ts +2 -0
  17. package/dist/gff/parseGFF.js +41 -0
  18. package/dist/gff/parseGFF.js.map +1 -0
  19. package/dist/gff/parseGFF.test.d.ts +1 -0
  20. package/dist/gff/parseGFF.test.js +92 -0
  21. package/dist/gff/parseGFF.test.js.map +1 -0
  22. package/dist/index.d.ts +5 -0
  23. package/dist/index.js +9 -0
  24. package/dist/index.js.map +1 -0
  25. package/dist/msa/A3mMSA.d.ts +33 -0
  26. package/dist/msa/A3mMSA.js +280 -0
  27. package/dist/msa/A3mMSA.js.map +1 -0
  28. package/dist/msa/A3mMSA.test.d.ts +1 -0
  29. package/dist/msa/A3mMSA.test.js +155 -0
  30. package/dist/msa/A3mMSA.test.js.map +1 -0
  31. package/dist/msa/ClustalMSA.d.ts +30 -0
  32. package/dist/msa/ClustalMSA.js +53 -0
  33. package/dist/msa/ClustalMSA.js.map +1 -0
  34. package/dist/msa/EmfMSA.d.ts +27 -0
  35. package/dist/msa/EmfMSA.js +53 -0
  36. package/dist/msa/EmfMSA.js.map +1 -0
  37. package/dist/msa/FastaMSA.d.ts +19 -0
  38. package/dist/msa/FastaMSA.js +69 -0
  39. package/dist/msa/FastaMSA.js.map +1 -0
  40. package/dist/msa/StockholmMSA.d.ts +54 -0
  41. package/dist/msa/StockholmMSA.js +113 -0
  42. package/dist/msa/StockholmMSA.js.map +1 -0
  43. package/dist/msa/index.d.ts +18 -0
  44. package/dist/msa/index.js +34 -0
  45. package/dist/msa/index.js.map +1 -0
  46. package/dist/msa/index.test.d.ts +1 -0
  47. package/dist/msa/index.test.js +60 -0
  48. package/dist/msa/index.test.js.map +1 -0
  49. package/dist/msa/parseNewick.d.ts +60 -0
  50. package/dist/msa/parseNewick.js +95 -0
  51. package/dist/msa/parseNewick.js.map +1 -0
  52. package/dist/msa/stockholmParser.d.ts +22 -0
  53. package/dist/msa/stockholmParser.js +141 -0
  54. package/dist/msa/stockholmParser.js.map +1 -0
  55. package/dist/msa/stockholmParser.test.d.ts +1 -0
  56. package/dist/msa/stockholmParser.test.js +111 -0
  57. package/dist/msa/stockholmParser.test.js.map +1 -0
  58. package/dist/types.d.ts +66 -0
  59. package/dist/types.js +2 -0
  60. package/dist/types.js.map +1 -0
  61. package/dist/util.d.ts +2 -0
  62. package/dist/util.js +10 -0
  63. package/dist/util.js.map +1 -0
  64. package/package.json +25 -0
  65. package/src/gff/gffToInterPro.test.ts +202 -0
  66. package/src/gff/gffToInterPro.ts +113 -0
  67. package/src/gff/index.ts +3 -0
  68. package/src/gff/interProToGFF.test.ts +206 -0
  69. package/src/gff/interProToGFF.ts +59 -0
  70. package/src/gff/parseGFF.test.ts +106 -0
  71. package/src/gff/parseGFF.ts +46 -0
  72. package/src/index.ts +29 -0
  73. package/src/msa/A3mMSA.test.ts +192 -0
  74. package/src/msa/A3mMSA.ts +320 -0
  75. package/src/msa/ClustalMSA.ts +67 -0
  76. package/src/msa/EmfMSA.ts +67 -0
  77. package/src/msa/FastaMSA.ts +82 -0
  78. package/src/msa/StockholmMSA.ts +141 -0
  79. package/src/msa/index.test.ts +74 -0
  80. package/src/msa/index.ts +44 -0
  81. package/src/msa/parseNewick.ts +94 -0
  82. package/src/msa/stockholmParser.test.ts +123 -0
  83. package/src/msa/stockholmParser.ts +157 -0
  84. package/src/types.ts +68 -0
  85. package/src/util.ts +19 -0
@@ -0,0 +1,95 @@
1
+ /**
2
+ * Newick format parser in JavaScript.
3
+ *
4
+ * Copyright (c) Jason Davies 2010.
5
+ *
6
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
7
+ * of this software and associated documentation files (the "Software"), to deal
8
+ * in the Software without restriction, including without limitation the rights
9
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
10
+ * copies of the Software, and to permit persons to whom the Software is
11
+ * furnished to do so, subject to the following conditions:
12
+ *
13
+ * The above copyright notice and this permission notice shall be included in
14
+ * all copies or substantial portions of the Software.
15
+ *
16
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
22
+ * THE SOFTWARE.
23
+ *
24
+ * Example tree (from http://en.wikipedia.org/wiki/Newick_format):
25
+ *
26
+ * +--0.1--A
27
+ * F-----0.2-----B +-------0.3----C
28
+ * +------------------0.5-----E
29
+ * +---------0.4------D
30
+ *
31
+ * Newick format:
32
+ * (A:0.1,B:0.2,(C:0.3,D:0.4)E:0.5)F;
33
+ *
34
+ * Converted to JSON:
35
+ * {
36
+ * name: "F",
37
+ * children: [
38
+ * {name: "A", length: 0.1},
39
+ * {name: "B", length: 0.2},
40
+ * {
41
+ * name: "E",
42
+ * length: 0.5,
43
+ * children: [
44
+ * {name: "C", length: 0.3},
45
+ * {name: "D", length: 0.4}
46
+ * ]
47
+ * }
48
+ * ]
49
+ * }
50
+ *
51
+ * Converted to JSON, but with no names or lengths:
52
+ * {
53
+ * children: [
54
+ * {}, {}, {
55
+ * children: [{}, {}]
56
+ * }
57
+ * ]
58
+ * }
59
+ */
60
+ export default function parse(s) {
61
+ const ancestors = [];
62
+ let tree = {};
63
+ const tokens = s.split(/\s*(;|\(|\)|,|:)\s*/);
64
+ for (let i = 0; i < tokens.length; i++) {
65
+ const token = tokens[i];
66
+ const subtree = {};
67
+ switch (token) {
68
+ case '(': // new children
69
+ tree.children = [subtree];
70
+ ancestors.push(tree);
71
+ tree = subtree;
72
+ break;
73
+ case ',': // another branch
74
+ ancestors.at(-1)?.children.push(subtree);
75
+ tree = subtree;
76
+ break;
77
+ case ')': // optional name next
78
+ tree = ancestors.pop();
79
+ break;
80
+ case ':': // optional length next
81
+ break;
82
+ default: {
83
+ const x = tokens[i - 1];
84
+ if (x === ')' || x === '(' || x === ',') {
85
+ tree.name = token;
86
+ }
87
+ else if (x === ':') {
88
+ tree.length = Number.parseFloat(token);
89
+ }
90
+ }
91
+ }
92
+ }
93
+ return tree;
94
+ }
95
+ //# sourceMappingURL=parseNewick.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"parseNewick.js","sourceRoot":"","sources":["../../src/msa/parseNewick.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA0DG;AACH,MAAM,CAAC,OAAO,UAAU,KAAK,CAAC,CAAS;IACrC,MAAM,SAAS,GAAG,EAAE,CAAA;IAEpB,IAAI,IAAI,GAAG,EAAyB,CAAA;IACpC,MAAM,MAAM,GAAG,CAAC,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAA;IAC7C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;QACxB,MAAM,OAAO,GAAG,EAAE,CAAA;QAClB,QAAQ,KAAK,EAAE,CAAC;YACd,KAAK,GAAG,EAAE,eAAe;gBACvB,IAAI,CAAC,QAAQ,GAAG,CAAC,OAAO,CAAC,CAAA;gBACzB,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;gBACpB,IAAI,GAAG,OAAO,CAAA;gBACd,MAAK;YACP,KAAK,GAAG,EAAE,iBAAiB;gBACzB,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;gBACxC,IAAI,GAAG,OAAO,CAAA;gBACd,MAAK;YACP,KAAK,GAAG,EAAE,qBAAqB;gBAC7B,IAAI,GAAG,SAAS,CAAC,GAAG,EAAG,CAAA;gBACvB,MAAK;YACP,KAAK,GAAG,EAAE,uBAAuB;gBAC/B,MAAK;YACP,OAAO,CAAC,CAAC,CAAC;gBACR,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;gBACxB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,GAAG,EAAE,CAAC;oBACxC,IAAI,CAAC,IAAI,GAAG,KAAK,CAAA;gBACnB,CAAC;qBAAM,IAAI,CAAC,KAAK,GAAG,EAAE,CAAC;oBACrB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,UAAU,CAAC,KAAK,CAAC,CAAA;gBACxC,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC"}
@@ -0,0 +1,22 @@
1
+ /**
2
+ * Stockholm format parser
3
+ * Based on stockholm-js by Ian Holmes
4
+ * Rewritten in TypeScript
5
+ */
6
+ export interface StockholmData {
7
+ gf: Record<string, string[]>;
8
+ gc: Record<string, string>;
9
+ gs: Record<string, Record<string, string[]>>;
10
+ gr: Record<string, Record<string, string>>;
11
+ seqname: string[];
12
+ seqdata: Record<string, string>;
13
+ }
14
+ export declare function sniff(text: string): boolean;
15
+ export declare function parseAll(text: string, opts?: {
16
+ strict?: boolean;
17
+ quiet?: boolean;
18
+ }): StockholmData[];
19
+ export declare function parse(text: string, opts?: {
20
+ strict?: boolean;
21
+ quiet?: boolean;
22
+ }): StockholmData;
@@ -0,0 +1,141 @@
1
+ /**
2
+ * Stockholm format parser
3
+ * Based on stockholm-js by Ian Holmes
4
+ * Rewritten in TypeScript
5
+ */
6
+ function createStockholm() {
7
+ return {
8
+ gf: {},
9
+ gc: {},
10
+ gs: {},
11
+ gr: {},
12
+ seqname: [],
13
+ seqdata: {},
14
+ };
15
+ }
16
+ const formatStartRegex = /^# STOCKHOLM 1.0/;
17
+ const formatEndRegex = /^\/\/\s*$/;
18
+ const gfRegex = /^#=GF\s+(\S+)\s+(.*?)\s*$/;
19
+ const gcRegex = /^#=GC\s+(\S+)\s+(.*?)\s*$/;
20
+ const gsRegex = /^#=GS\s+(\S+)\s+(\S+)\s+(.*?)\s*$/;
21
+ const grRegex = /^#=GR\s+(\S+)\s+(\S+)\s+(.*?)\s*$/;
22
+ const lineRegex = /^\s*(\S+)\s+(\S+)\s*$/;
23
+ const nonwhiteRegex = /\S/;
24
+ export function sniff(text) {
25
+ return formatStartRegex.test(text);
26
+ }
27
+ export function parseAll(text, opts) {
28
+ const options = opts ?? {};
29
+ const db = [];
30
+ let stock = null;
31
+ const lines = text.split('\n');
32
+ for (const line of lines) {
33
+ let match;
34
+ if (formatStartRegex.test(line)) {
35
+ stock = createStockholm();
36
+ }
37
+ else if (formatEndRegex.test(line)) {
38
+ if (stock) {
39
+ db.push(stock);
40
+ }
41
+ stock = null;
42
+ }
43
+ else if ((match = gfRegex.exec(line))) {
44
+ if (!stock) {
45
+ if (options.strict) {
46
+ throw new Error('No format header: # STOCKHOLM 1.0');
47
+ }
48
+ stock = createStockholm();
49
+ }
50
+ const tag = match[1];
51
+ if (!stock.gf[tag]) {
52
+ stock.gf[tag] = [];
53
+ }
54
+ stock.gf[tag].push(match[2]);
55
+ }
56
+ else if ((match = gcRegex.exec(line))) {
57
+ if (!stock) {
58
+ if (options.strict) {
59
+ throw new Error('No format header: # STOCKHOLM 1.0');
60
+ }
61
+ stock = createStockholm();
62
+ }
63
+ const tag = match[1];
64
+ if (!stock.gc[tag]) {
65
+ stock.gc[tag] = '';
66
+ }
67
+ stock.gc[tag] += match[2];
68
+ }
69
+ else if ((match = gsRegex.exec(line))) {
70
+ if (!stock) {
71
+ if (options.strict) {
72
+ throw new Error('No format header: # STOCKHOLM 1.0');
73
+ }
74
+ stock = createStockholm();
75
+ }
76
+ const seqname = match[1];
77
+ const tag = match[2];
78
+ const value = match[3];
79
+ if (!stock.gs[tag]) {
80
+ stock.gs[tag] = {};
81
+ }
82
+ if (!stock.gs[tag][seqname]) {
83
+ stock.gs[tag][seqname] = [];
84
+ }
85
+ stock.gs[tag][seqname].push(value);
86
+ }
87
+ else if ((match = grRegex.exec(line))) {
88
+ if (!stock) {
89
+ if (options.strict) {
90
+ throw new Error('No format header: # STOCKHOLM 1.0');
91
+ }
92
+ stock = createStockholm();
93
+ }
94
+ const seqname = match[1];
95
+ const tag = match[2];
96
+ const value = match[3];
97
+ if (!stock.gr[tag]) {
98
+ stock.gr[tag] = {};
99
+ }
100
+ if (!stock.gr[tag][seqname]) {
101
+ stock.gr[tag][seqname] = '';
102
+ }
103
+ stock.gr[tag][seqname] += value;
104
+ }
105
+ else if ((match = lineRegex.exec(line))) {
106
+ if (!stock) {
107
+ if (options.strict) {
108
+ throw new Error('No format header: # STOCKHOLM 1.0');
109
+ }
110
+ stock = createStockholm();
111
+ }
112
+ const seqname = match[1];
113
+ const seqdata = match[2];
114
+ if (!stock.seqdata[seqname]) {
115
+ stock.seqdata[seqname] = '';
116
+ stock.seqname.push(seqname);
117
+ }
118
+ stock.seqdata[seqname] += seqdata;
119
+ }
120
+ else if (nonwhiteRegex.test(line)) {
121
+ if (options.strict) {
122
+ throw new Error('Malformed line');
123
+ }
124
+ }
125
+ }
126
+ if (stock) {
127
+ db.push(stock);
128
+ }
129
+ return db;
130
+ }
131
+ export function parse(text, opts) {
132
+ const db = parseAll(text, opts);
133
+ if (db.length === 0) {
134
+ throw new Error('No alignments found');
135
+ }
136
+ if (db.length > 1) {
137
+ throw new Error('More than one alignment found');
138
+ }
139
+ return db[0];
140
+ }
141
+ //# sourceMappingURL=stockholmParser.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"stockholmParser.js","sourceRoot":"","sources":["../../src/msa/stockholmParser.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAWH,SAAS,eAAe;IACtB,OAAO;QACL,EAAE,EAAE,EAAE;QACN,EAAE,EAAE,EAAE;QACN,EAAE,EAAE,EAAE;QACN,EAAE,EAAE,EAAE;QACN,OAAO,EAAE,EAAE;QACX,OAAO,EAAE,EAAE;KACZ,CAAA;AACH,CAAC;AAED,MAAM,gBAAgB,GAAG,kBAAkB,CAAA;AAC3C,MAAM,cAAc,GAAG,WAAW,CAAA;AAClC,MAAM,OAAO,GAAG,2BAA2B,CAAA;AAC3C,MAAM,OAAO,GAAG,2BAA2B,CAAA;AAC3C,MAAM,OAAO,GAAG,mCAAmC,CAAA;AACnD,MAAM,OAAO,GAAG,mCAAmC,CAAA;AACnD,MAAM,SAAS,GAAG,uBAAuB,CAAA;AACzC,MAAM,aAAa,GAAG,IAAI,CAAA;AAE1B,MAAM,UAAU,KAAK,CAAC,IAAY;IAChC,OAAO,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACpC,CAAC;AAED,MAAM,UAAU,QAAQ,CACtB,IAAY,EACZ,IAA4C;IAE5C,MAAM,OAAO,GAAG,IAAI,IAAI,EAAE,CAAA;IAC1B,MAAM,EAAE,GAAoB,EAAE,CAAA;IAC9B,IAAI,KAAK,GAAyB,IAAI,CAAA;IAEtC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;IAC9B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,KAA6B,CAAA;QAEjC,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAChC,KAAK,GAAG,eAAe,EAAE,CAAA;QAC3B,CAAC;aAAM,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACrC,IAAI,KAAK,EAAE,CAAC;gBACV,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YAChB,CAAC;YACD,KAAK,GAAG,IAAI,CAAA;QACd,CAAC;aAAM,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;oBACnB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAA;gBACtD,CAAC;gBACD,KAAK,GAAG,eAAe,EAAE,CAAA;YAC3B,CAAC;YACD,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACrB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC;gBACnB,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,EAAE,CAAA;YACpB,CAAC;YACD,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,CAAA;QAC/B,CAAC;aAAM,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;oBACnB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAA;gBACtD,CAAC;gBACD,KAAK,GAAG,eAAe,EAAE,CAAA;YAC3B,CAAC;YACD,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACrB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC;gBACnB,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,EAAE,CAAA;YACpB,CAAC;YACD,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,CAAC,CAAE,CAAA;QAC5B,CAAC;aAAM,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;oBACnB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAA;gBACtD,CAAC;gBACD,KAAK,GAAG,eAAe,EAAE,CAAA;YAC3B,CAAC;YACD,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACzB,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACrB,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACvB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC;gBACnB,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,EAAE,CAAA;YACpB,CAAC;YACD,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC5B,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,CAAA;YAC7B,CAAC;YACD,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACpC,CAAC;aAAM,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;YACxC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;oBACnB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAA;gBACtD,CAAC;gBACD,KAAK,GAAG,eAAe,EAAE,CAAA;YAC3B,CAAC;YACD,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACzB,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACrB,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACvB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC;gBACnB,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,EAAE,CAAA;YACpB,CAAC;YACD,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC5B,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,CAAA;YAC7B,CAAC;YACD,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,IAAI,KAAK,CAAA;QACjC,CAAC;aAAM,IAAI,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;YAC1C,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;oBACnB,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAA;gBACtD,CAAC;gBACD,KAAK,GAAG,eAAe,EAAE,CAAA;YAC3B,CAAC;YACD,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACzB,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;YACzB,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC5B,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,GAAG,EAAE,CAAA;gBAC3B,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;YAC7B,CAAC;YACD,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,OAAO,CAAA;QACnC,CAAC;aAAM,IAAI,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACpC,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;gBACnB,MAAM,IAAI,KAAK,CAAC,gBAAgB,CAAC,CAAA;YACnC,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,KAAK,EAAE,CAAC;QACV,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;IAChB,CAAC;IAED,OAAO,EAAE,CAAA;AACX,CAAC;AAED,MAAM,UAAU,KAAK,CACnB,IAAY,EACZ,IAA4C;IAE5C,MAAM,EAAE,GAAG,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;IAC/B,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACpB,MAAM,IAAI,KAAK,CAAC,qBAAqB,CAAC,CAAA;IACxC,CAAC;IACD,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClB,MAAM,IAAI,KAAK,CAAC,+BAA+B,CAAC,CAAA;IAClD,CAAC;IACD,OAAO,EAAE,CAAC,CAAC,CAAE,CAAA;AACf,CAAC"}
@@ -0,0 +1 @@
1
+ export {};
@@ -0,0 +1,111 @@
1
+ import { describe, expect, test } from 'vitest';
2
+ import { parse, parseAll, sniff } from './stockholmParser';
3
+ describe('stockholmParser', () => {
4
+ describe('sniff', () => {
5
+ test('returns true for Stockholm format', () => {
6
+ expect(sniff('# STOCKHOLM 1.0\n')).toBe(true);
7
+ });
8
+ test('returns false for non-Stockholm format', () => {
9
+ expect(sniff('>seq1\nACDEF')).toBe(false);
10
+ expect(sniff('CLUSTAL W')).toBe(false);
11
+ });
12
+ });
13
+ describe('parseAll', () => {
14
+ test('parses simple Stockholm file', () => {
15
+ const stockholm = `# STOCKHOLM 1.0
16
+ seq1 ACDEFGHIKL
17
+ seq2 ACDEFGHIKL
18
+ //`;
19
+ const result = parseAll(stockholm);
20
+ expect(result).toHaveLength(1);
21
+ expect(result[0]?.seqdata).toEqual({
22
+ seq1: 'ACDEFGHIKL',
23
+ seq2: 'ACDEFGHIKL',
24
+ });
25
+ expect(result[0]?.seqname).toEqual(['seq1', 'seq2']);
26
+ });
27
+ test('parses GF lines', () => {
28
+ const stockholm = `# STOCKHOLM 1.0
29
+ #=GF DE Description of the alignment
30
+ #=GF NH (A:0.1,B:0.2);
31
+ seq1 ACDEF
32
+ //`;
33
+ const result = parseAll(stockholm);
34
+ expect(result[0]?.gf.DE).toEqual(['Description of the alignment']);
35
+ expect(result[0]?.gf.NH).toEqual(['(A:0.1,B:0.2);']);
36
+ });
37
+ test('parses GC lines', () => {
38
+ const stockholm = `# STOCKHOLM 1.0
39
+ seq1 ACDEF
40
+ #=GC seq_cons *****
41
+ #=GC SS_cons <<>>>
42
+ //`;
43
+ const result = parseAll(stockholm);
44
+ expect(result[0]?.gc.seq_cons).toBe('*****');
45
+ expect(result[0]?.gc.SS_cons).toBe('<<>>>');
46
+ });
47
+ test('parses GS lines', () => {
48
+ const stockholm = `# STOCKHOLM 1.0
49
+ #=GS seq1 AC P12345
50
+ #=GS seq1 DR PDB; 1ABC A; 1-100
51
+ seq1 ACDEF
52
+ //`;
53
+ const result = parseAll(stockholm);
54
+ expect(result[0]?.gs.AC?.seq1).toEqual(['P12345']);
55
+ expect(result[0]?.gs.DR?.seq1).toEqual(['PDB; 1ABC A; 1-100']);
56
+ });
57
+ test('parses multiple alignments', () => {
58
+ const stockholm = `# STOCKHOLM 1.0
59
+ seq1 ACDEF
60
+ //
61
+ # STOCKHOLM 1.0
62
+ seq2 GHIKL
63
+ //`;
64
+ const result = parseAll(stockholm);
65
+ expect(result).toHaveLength(2);
66
+ expect(result[0]?.seqdata.seq1).toBe('ACDEF');
67
+ expect(result[1]?.seqdata.seq2).toBe('GHIKL');
68
+ });
69
+ test('concatenates interleaved sequence blocks', () => {
70
+ const stockholm = `# STOCKHOLM 1.0
71
+ seq1 ACDEF
72
+ seq2 GHIKL
73
+
74
+ seq1 MNOPQ
75
+ seq2 RSTUV
76
+ //`;
77
+ const result = parseAll(stockholm);
78
+ expect(result[0]?.seqdata.seq1).toBe('ACDEFMNOPQ');
79
+ expect(result[0]?.seqdata.seq2).toBe('GHIKLRSTUV');
80
+ });
81
+ test('handles file without trailing //', () => {
82
+ const stockholm = `# STOCKHOLM 1.0
83
+ seq1 ACDEF`;
84
+ const result = parseAll(stockholm);
85
+ expect(result).toHaveLength(1);
86
+ expect(result[0]?.seqdata.seq1).toBe('ACDEF');
87
+ });
88
+ });
89
+ describe('parse', () => {
90
+ test('returns single alignment', () => {
91
+ const stockholm = `# STOCKHOLM 1.0
92
+ seq1 ACDEF
93
+ //`;
94
+ const result = parse(stockholm);
95
+ expect(result.seqdata.seq1).toBe('ACDEF');
96
+ });
97
+ test('throws on empty input', () => {
98
+ expect(() => parse('')).toThrow('No alignments found');
99
+ });
100
+ test('throws on multiple alignments', () => {
101
+ const stockholm = `# STOCKHOLM 1.0
102
+ seq1 ACDEF
103
+ //
104
+ # STOCKHOLM 1.0
105
+ seq2 GHIKL
106
+ //`;
107
+ expect(() => parse(stockholm)).toThrow('More than one alignment found');
108
+ });
109
+ });
110
+ });
111
+ //# sourceMappingURL=stockholmParser.test.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"stockholmParser.test.js","sourceRoot":"","sources":["../../src/msa/stockholmParser.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,QAAQ,CAAA;AAE/C,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,MAAM,mBAAmB,CAAA;AAE1D,QAAQ,CAAC,iBAAiB,EAAE,GAAG,EAAE;IAC/B,QAAQ,CAAC,OAAO,EAAE,GAAG,EAAE;QACrB,IAAI,CAAC,mCAAmC,EAAE,GAAG,EAAE;YAC7C,MAAM,CAAC,KAAK,CAAC,mBAAmB,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAC/C,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,wCAAwC,EAAE,GAAG,EAAE;YAClD,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACzC,MAAM,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACxC,CAAC,CAAC,CAAA;IACJ,CAAC,CAAC,CAAA;IAEF,QAAQ,CAAC,UAAU,EAAE,GAAG,EAAE;QACxB,IAAI,CAAC,8BAA8B,EAAE,GAAG,EAAE;YACxC,MAAM,SAAS,GAAG;;;GAGrB,CAAA;YACG,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAA;YAC9B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,OAAO,CAAC;gBACjC,IAAI,EAAE,YAAY;gBAClB,IAAI,EAAE,YAAY;aACnB,CAAC,CAAA;YACF,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAAA;QACtD,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,iBAAiB,EAAE,GAAG,EAAE;YAC3B,MAAM,SAAS,GAAG;;;;GAIrB,CAAA;YACG,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,8BAA8B,CAAC,CAAC,CAAA;YAClE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,gBAAgB,CAAC,CAAC,CAAA;QACtD,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,iBAAiB,EAAE,GAAG,EAAE;YAC3B,MAAM,SAAS,GAAG;;;;GAIrB,CAAA;YACG,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;YAC5C,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAC7C,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,iBAAiB,EAAE,GAAG,EAAE;YAC3B,MAAM,SAAS,GAAG;;;;GAIrB,CAAA;YACG,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAA;YAClD,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,oBAAoB,CAAC,CAAC,CAAA;QAChE,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,4BAA4B,EAAE,GAAG,EAAE;YACtC,MAAM,SAAS,GAAG;;;;;GAKrB,CAAA;YACG,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAA;YAC9B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;YAC7C,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAC/C,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,0CAA0C,EAAE,GAAG,EAAE;YACpD,MAAM,SAAS,GAAG;;;;;;GAMrB,CAAA;YACG,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAA;YAClD,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAA;QACpD,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,kCAAkC,EAAE,GAAG,EAAE;YAC5C,MAAM,SAAS,GAAG;YACZ,CAAA;YACN,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;YAClC,MAAM,CAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAA;YAC9B,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAC/C,CAAC,CAAC,CAAA;IACJ,CAAC,CAAC,CAAA;IAEF,QAAQ,CAAC,OAAO,EAAE,GAAG,EAAE;QACrB,IAAI,CAAC,0BAA0B,EAAE,GAAG,EAAE;YACpC,MAAM,SAAS,GAAG;;GAErB,CAAA;YACG,MAAM,MAAM,GAAG,KAAK,CAAC,SAAS,CAAC,CAAA;YAC/B,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;QAC3C,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,uBAAuB,EAAE,GAAG,EAAE;YACjC,MAAM,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,qBAAqB,CAAC,CAAA;QACxD,CAAC,CAAC,CAAA;QAEF,IAAI,CAAC,+BAA+B,EAAE,GAAG,EAAE;YACzC,MAAM,SAAS,GAAG;;;;;GAKrB,CAAA;YACG,MAAM,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,+BAA+B,CAAC,CAAA;QACzE,CAAC,CAAC,CAAA;IACJ,CAAC,CAAC,CAAA;AACJ,CAAC,CAAC,CAAA"}
@@ -0,0 +1,66 @@
1
+ export interface Node {
2
+ children?: Node[];
3
+ name?: string;
4
+ length?: number;
5
+ }
6
+ export interface NodeWithIds {
7
+ id: string;
8
+ name: string;
9
+ children: NodeWithIds[];
10
+ length?: number;
11
+ noTree?: boolean;
12
+ }
13
+ export interface NodeWithIdsAndLength {
14
+ id: string;
15
+ name: string;
16
+ children: NodeWithIdsAndLength[];
17
+ noTree?: boolean;
18
+ length: number;
19
+ }
20
+ export interface GFFRecord {
21
+ seq_id: string;
22
+ source: string;
23
+ type: string;
24
+ start: number;
25
+ end: number;
26
+ score: number;
27
+ strand: string;
28
+ phase: string;
29
+ [key: string]: string | number;
30
+ }
31
+ export interface InterProScanMatch {
32
+ signature: {
33
+ entry?: {
34
+ name: string;
35
+ description: string;
36
+ accession: string;
37
+ };
38
+ };
39
+ locations: {
40
+ start: number;
41
+ end: number;
42
+ }[];
43
+ }
44
+ export interface InterProScanResults {
45
+ matches: InterProScanMatch[];
46
+ xref: {
47
+ id: string;
48
+ }[];
49
+ }
50
+ export interface InterProScanResponse {
51
+ results: InterProScanResults[];
52
+ }
53
+ export interface MSAParser {
54
+ getMSA(): unknown;
55
+ getRow(name: string): string;
56
+ getWidth(): number;
57
+ getNames(): string[];
58
+ getTree(): NodeWithIds;
59
+ getHeader(): unknown;
60
+ getRowData(name: string): unknown;
61
+ getStructures(): Record<string, unknown>;
62
+ alignmentNames: string[];
63
+ seqConsensus: string | undefined;
64
+ secondaryStructureConsensus: string | undefined;
65
+ tracks: unknown[];
66
+ }
package/dist/types.js ADDED
@@ -0,0 +1,2 @@
1
+ export {};
2
+ //# sourceMappingURL=types.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"types.js","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":""}
package/dist/util.d.ts ADDED
@@ -0,0 +1,2 @@
1
+ import type { Node, NodeWithIds } from './types';
2
+ export declare function generateNodeIds(tree: Node, parent?: string, depth?: number): NodeWithIds;
package/dist/util.js ADDED
@@ -0,0 +1,10 @@
1
+ export function generateNodeIds(tree, parent = 'node', depth = 0) {
2
+ const id = `${parent}-${depth}`;
3
+ return {
4
+ ...tree,
5
+ id,
6
+ name: tree.name || id,
7
+ children: tree.children?.map((b, i) => generateNodeIds(b, `${id}-${i}`, depth + 1)) || [],
8
+ };
9
+ }
10
+ //# sourceMappingURL=util.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"util.js","sourceRoot":"","sources":["../src/util.ts"],"names":[],"mappings":"AAEA,MAAM,UAAU,eAAe,CAC7B,IAAU,EACV,MAAM,GAAG,MAAM,EACf,KAAK,GAAG,CAAC;IAET,MAAM,EAAE,GAAG,GAAG,MAAM,IAAI,KAAK,EAAE,CAAA;IAE/B,OAAO;QACL,GAAG,IAAI;QACP,EAAE;QACF,IAAI,EAAE,IAAI,CAAC,IAAI,IAAI,EAAE;QACrB,QAAQ,EACN,IAAI,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAC1B,eAAe,CAAC,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,EAAE,EAAE,KAAK,GAAG,CAAC,CAAC,CAC5C,IAAI,EAAE;KACV,CAAA;AACH,CAAC"}
package/package.json ADDED
@@ -0,0 +1,25 @@
1
+ {
2
+ "name": "msa-parsers",
3
+ "version": "1.0.3",
4
+ "license": "MIT",
5
+ "main": "dist/index.js",
6
+ "module": "dist/index.js",
7
+ "types": "dist/index.d.ts",
8
+ "files": [
9
+ "dist",
10
+ "src"
11
+ ],
12
+ "scripts": {
13
+ "clean": "rimraf dist",
14
+ "build": "tsc",
15
+ "watch": "tsc --watch",
16
+ "test": "vitest"
17
+ },
18
+ "dependencies": {
19
+ "clustal-js": "^2.0.13",
20
+ "emf-js": "^2.0.0"
21
+ },
22
+ "publishConfig": {
23
+ "access": "public"
24
+ }
25
+ }