p2s_mapper 1.4.0 → 2.0.0

package/esm/pdbeBestStructures.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"pdbeBestStructures.js","sourceRoot":"","sources":["../src/pdbeBestStructures.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,4EAA4E;AAC5E,+EAA+E;AAe/E,MAAM,UAAU,qBAAqB,CAAC,SAAiB;IACrD,OAAO,2DAA2D,SAAS,EAAE,CAAA;AAC/E,CAAC;AAED,6EAA6E;AAC7E,8EAA8E;AAC9E,gBAAgB;AAChB,MAAM,UAAU,OAAO,CAAC,KAAa;IACnC,OAAO,uBAAuB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;AAC5C,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,KAAa;IACxC,OAAO,kCAAkC,KAAK,CAAC,WAAW,EAAE,EAAE,CAAA;AAChE,CAAC;AAED,SAAS,QAAQ,CAAC,KAAc;IAC9B,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAA;AAC7E,CAAC;AAED,SAAS,YAAY,CAAC,KAAc;IAClC,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAA;AAChF,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,mBAAmB,CAAC,IAAa;IAC/C,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAA;IAChE,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC;QACzB,OAAO,EAAE,CAAA;IACX,CAAC;IACD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA6B,CAAA;IACjD,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;YACnB,SAAQ;QACV,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,GAAG,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAA;QACrE,MAAM,KAAK,GAAG,OAAO,GAAG,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAA;QACzE,MAAM,QAAQ,GAAG,YAAY,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;QAC5C,MAAM,MAAM,GAAG,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;QACxC,MAAM,QAAQ,GAAG,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;QAC3C,IACE,KAAK,KAAK,SAAS;YACnB,KAAK,KAAK,SAAS;YACnB,QAAQ,KAAK,SAAS;YACtB,MAAM,KAAK,SAAS;YACpB,QAAQ,KAAK,SAAS,EACtB,CAAC;YACD,SAAQ;QACV,CAAC;QACD,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAA;QAChC,IAAI,QAAQ,EAAE,CAAC;YACb,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;gBACrC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YAC7B,CAAC;YACD,SAAQ;QACV,CAAC;QACD,MAAM,UAAU,GAAG,YAAY,CAAC,GAAG,CAAC,UAAU,CAAC,CAAA;QAC/C,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE;YACd,KAAK;YACL,kBAAkB,EAChB,OAAO,GAAG,CAAC,mBAAmB,KAAK,QAAQ;gBACzC,CAAC,CAAC,GAAG,CAAC,mBAAmB;gBACzB,CAAC,CAAC,SAAS;YACf,GAAG,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC;YACnD,QAAQ;YACR,MAAM;YACN,QAAQ;YACR,MAAM,EAAE,CAAC,KAAK,CAAC;SAChB,CAAC,CAAA;IACJ,CAAC;IACD,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;AAC3B,CAAC"}

package/esm/pdbeEntryMolecules.d.ts

@@ -1,7 +0,0 @@
-export declare function pdbeEntryMoleculesUrl(pdbId: string): string;
-/**
- * One sequence per protein entity, in entity order. Nucleic-acid entities are
- * left out: A, C, G, T and U are amino-acid letters too, so a DNA chain scores
- * as a candidate rather than losing honestly.
- */
-export declare function parseEntryMolecules(json: unknown): string[];

package/esm/pdbeEntryMolecules.js

@@ -1,35 +0,0 @@
-// The polymer entities of a PDB entry, as PDBe already knows them. The dialog
-// wants the entry's residues only to annotate the isoform list, and reading
-// them out of the structure file meant downloading and parsing the whole
-// mmCIF — minutes for a ribosome, for a label. This is a few kB of JSON, and
-// the view downloads the file itself when the user actually launches.
-export function pdbeEntryMoleculesUrl(pdbId) {
-    return `https://www.ebi.ac.uk/pdbe/api/pdb/entry/molecules/${pdbId.toLowerCase()}`;
-}
-function isRecord(value) {
-    return typeof value === 'object' && value !== null && !Array.isArray(value);
-}
-/**
- * One sequence per protein entity, in entity order. Nucleic-acid entities are
- * left out: A, C, G, T and U are amino-acid letters too, so a DNA chain scores
- * as a candidate rather than losing honestly.
- */
-export function parseEntryMolecules(json) {
-    const entities = isRecord(json) ? Object.values(json)[0] : undefined;
-    if (!Array.isArray(entities)) {
-        return [];
-    }
-    return entities.flatMap(entity => {
-        if (!isRecord(entity)) {
-            return [];
-        }
-        const { molecule_type: type, sequence } = entity;
-        return typeof type === 'string' &&
-            type.startsWith('polypeptide') &&
-            typeof sequence === 'string' &&
-            sequence.length > 0
-            ? [sequence]
-            : [];
-    });
-}
-//# sourceMappingURL=pdbeEntryMolecules.js.map

package/esm/pdbeEntryMolecules.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"pdbeEntryMolecules.js","sourceRoot":"","sources":["../src/pdbeEntryMolecules.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,4EAA4E;AAC5E,yEAAyE;AACzE,6EAA6E;AAC7E,sEAAsE;AAEtE,MAAM,UAAU,qBAAqB,CAAC,KAAa;IACjD,OAAO,sDAAsD,KAAK,CAAC,WAAW,EAAE,EAAE,CAAA;AACpF,CAAC;AAED,SAAS,QAAQ,CAAC,KAAc;IAC9B,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAA;AAC7E,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,mBAAmB,CAAC,IAAa;IAC/C,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAA;IACpE,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7B,OAAO,EAAE,CAAA;IACX,CAAC;IACD,OAAO,QAAQ,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE;QAC/B,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;YACtB,OAAO,EAAE,CAAA;QACX,CAAC;QACD,MAAM,EAAE,aAAa,EAAE,IAAI,EAAE,QAAQ,EAAE,GAAG,MAAM,CAAA;QAChD,OAAO,OAAO,IAAI,KAAK,QAAQ;YAC7B,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC;YAC9B,OAAO,QAAQ,KAAK,QAAQ;YAC5B,QAAQ,CAAC,MAAM,GAAG,CAAC;YACnB,CAAC,CAAC,CAAC,QAAQ,CAAC;YACZ,CAAC,CAAC,EAAE,CAAA;IACR,CAAC,CAAC,CAAA;AACJ,CAAC"}

package/esm/siftsNumbering.d.ts

@@ -1,23 +0,0 @@
-import type { UniProtStructureSegment } from './pdbUniProtMapping.ts';
-export interface AuthorRange {
-    start: number;
-    end: number;
-    chain: string;
-    /** author number minus UniProt number in the chosen segment */
-    shift: number;
-}
-/** The segments SIFTS gives one accession, chain ids and author numbers
- * included, out of a whole entry's mappings. */
-export declare function segmentsForAccession(mappings: readonly {
-    accession: string;
-    segments: UniProtStructureSegment[];
-}[], accession: string): UniProtStructureSegment[];
-/**
- * The author-numbered range a UniProt range is cited by in the entry: the
- * segment overlapping most of the range decides the shift. Undefined when no
- * chain of the entry covers any of it, or when SIFTS named no author numbers.
- */
-export declare function toAuthorRange(segments: readonly UniProtStructureSegment[], range: {
-    start: number;
-    end: number;
-}): AuthorRange | undefined;

package/esm/siftsNumbering.js

@@ -1,40 +0,0 @@
-// How a PDB entry numbers the residues of a UniProt sequence. Author numbering
-// is what a paper cites and what a viewer lights a named residue by, and it is
-// not the UniProt numbering for every entry: haemoglobin's chains number from
-// the mature protein, so author residue 1 is UniProt residue 2 (2HHB, measured
-// 2026-09-12). SIFTS records the correspondence per chain segment, and
-// `fetchUniProtStructureMappings` is what asks PDBe for it.
-/** The segments SIFTS gives one accession, chain ids and author numbers
- * included, out of a whole entry's mappings. */
-export function segmentsForAccession(mappings, accession) {
-    return mappings.find(m => m.accession === accession)?.segments ?? [];
-}
-/**
- * The author-numbered range a UniProt range is cited by in the entry: the
- * segment overlapping most of the range decides the shift. Undefined when no
- * chain of the entry covers any of it, or when SIFTS named no author numbers.
- */
-export function toAuthorRange(segments, range) {
-    const overlap = (s) => Math.min(s.unpEnd, range.end) - Math.max(s.unpStart, range.start) + 1;
-    const best = segments
-        .flatMap(s => {
-        const { chainId, authorStart, unpStart, unpEnd } = s;
-        return chainId !== undefined &&
-            authorStart !== undefined &&
-            overlap(s) > 0
-            ? [{ chainId, authorStart, unpStart, unpEnd, overlap: overlap(s) }]
-            : [];
-    })
-        .toSorted((a, b) => b.overlap - a.overlap)[0];
-    if (!best) {
-        return undefined;
-    }
-    const shift = best.authorStart - best.unpStart;
-    return {
-        start: Math.max(range.start, best.unpStart) + shift,
-        end: Math.min(range.end, best.unpEnd) + shift,
-        chain: best.chainId,
-        shift,
-    };
-}
-//# sourceMappingURL=siftsNumbering.js.map

package/esm/siftsNumbering.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"siftsNumbering.js","sourceRoot":"","sources":["../src/siftsNumbering.ts"],"names":[],"mappings":"AAAA,+EAA+E;AAC/E,+EAA+E;AAC/E,8EAA8E;AAC9E,+EAA+E;AAC/E,uEAAuE;AACvE,4DAA4D;AAY5D;gDACgD;AAChD,MAAM,UAAU,oBAAoB,CAClC,QAGG,EACH,SAAiB;IAEjB,OAAO,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,KAAK,SAAS,CAAC,EAAE,QAAQ,IAAI,EAAE,CAAA;AACtE,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAC3B,QAA4C,EAC5C,KAAqC;IAErC,MAAM,OAAO,GAAG,CAAC,CAA0B,EAAE,EAAE,CAC7C,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,EAAE,KAAK,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,EAAE,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;IACvE,MAAM,IAAI,GAAG,QAAQ;SAClB,OAAO,CAAC,CAAC,CAAC,EAAE;QACX,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,CAAC,CAAA;QACpD,OAAO,OAAO,KAAK,SAAS;YAC1B,WAAW,KAAK,SAAS;YACzB,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC;YACd,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;YACnE,CAAC,CAAC,EAAE,CAAA;IACR,CAAC,CAAC;SACD,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAA;IAC/C,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,SAAS,CAAA;IAClB,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC,QAAQ,CAAA;IAC9C,OAAO;QACL,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,QAAQ,CAAC,GAAG,KAAK;QACnD,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,EAAE,IAAI,CAAC,MAAM,CAAC,GAAG,KAAK;QAC7C,KAAK,EAAE,IAAI,CAAC,OAAO;QACnB,KAAK;KACN,CAAA;AACH,CAAC"}

package/src/pdbeBestStructures.ts

@@ -1,94 +0,0 @@
-// PDBe's SIFTS-derived list of the experimental structures covering a UniProt
-// entry, one row per chain, ranked by PDBe (coverage, then resolution). The
-// dialog shows one row per entry, so the chains of a homomer collapse into it.
-
-export interface PdbStructureEntry {
-  pdbId: string
-  experimentalMethod: string
-  /** Å; absent for NMR and some EM entries */
-  resolution?: number
-  /** 1-based inclusive UniProt positions the entry covers */
-  unpStart: number
-  unpEnd: number
-  /** share of the UniProt sequence the entry covers, 0..1 */
-  coverage: number
-  chains: string[]
-}
-
-export function pdbeBestStructuresUrl(uniprotId: string) {
-  return `https://www.ebi.ac.uk/pdbe/api/mappings/best_structures/${uniprotId}`
-}
-
-// A PDB id is four characters, the first a digit: 1TUP, 6ZIO. Checked before
-// a typed one reaches a fetch, so three characters on the way to four are not
-// a 404 apiece.
-export function isPdbId(value: string) {
-  return /^[1-9][A-Za-z0-9]{3}$/.test(value)
-}
-
-export function rcsbEntryUrl(pdbId: string) {
-  return `https://www.rcsb.org/structure/${pdbId.toUpperCase()}`
-}
-
-function isRecord(value: unknown): value is Record<string, unknown> {
-  return typeof value === 'object' && value !== null && !Array.isArray(value)
-}
-
-function finiteNumber(value: unknown) {
-  return typeof value === 'number' && Number.isFinite(value) ? value : undefined
-}
-
-/**
- * Groups the per-chain rows into one entry per PDB id, keeping PDBe's order.
- * An entry whose chains cover different UniProt ranges (a construct
- * crystallised beside a different fragment of itself) keeps its first row's
- * range, which is the one PDBe ranked it by. The response is keyed by
- * accession, so the single entry is taken whatever its key.
- */
-export function parseBestStructures(json: unknown): PdbStructureEntry[] {
-  const rows = isRecord(json) ? Object.values(json)[0] : undefined
-  if (!Array.isArray(rows)) {
-    return []
-  }
-  const byId = new Map<string, PdbStructureEntry>()
-  for (const row of rows) {
-    if (!isRecord(row)) {
-      continue
-    }
-    const pdbId = typeof row.pdb_id === 'string' ? row.pdb_id : undefined
-    const chain = typeof row.chain_id === 'string' ? row.chain_id : undefined
-    const unpStart = finiteNumber(row.unp_start)
-    const unpEnd = finiteNumber(row.unp_end)
-    const coverage = finiteNumber(row.coverage)
-    if (
-      pdbId === undefined ||
-      chain === undefined ||
-      unpStart === undefined ||
-      unpEnd === undefined ||
-      coverage === undefined
-    ) {
-      continue
-    }
-    const existing = byId.get(pdbId)
-    if (existing) {
-      if (!existing.chains.includes(chain)) {
-        existing.chains.push(chain)
-      }
-      continue
-    }
-    const resolution = finiteNumber(row.resolution)
-    byId.set(pdbId, {
-      pdbId,
-      experimentalMethod:
-        typeof row.experimental_method === 'string'
-          ? row.experimental_method
-          : 'unknown',
-      ...(resolution === undefined ? {} : { resolution }),
-      unpStart,
-      unpEnd,
-      coverage,
-      chains: [chain],
-    })
-  }
-  return [...byId.values()]
-}

package/src/pdbeEntryMolecules.ts

@@ -1,37 +0,0 @@
-// The polymer entities of a PDB entry, as PDBe already knows them. The dialog
-// wants the entry's residues only to annotate the isoform list, and reading
-// them out of the structure file meant downloading and parsing the whole
-// mmCIF — minutes for a ribosome, for a label. This is a few kB of JSON, and
-// the view downloads the file itself when the user actually launches.
-
-export function pdbeEntryMoleculesUrl(pdbId: string) {
-  return `https://www.ebi.ac.uk/pdbe/api/pdb/entry/molecules/${pdbId.toLowerCase()}`
-}
-
-function isRecord(value: unknown): value is Record<string, unknown> {
-  return typeof value === 'object' && value !== null && !Array.isArray(value)
-}
-
-/**
- * One sequence per protein entity, in entity order. Nucleic-acid entities are
- * left out: A, C, G, T and U are amino-acid letters too, so a DNA chain scores
- * as a candidate rather than losing honestly.
- */
-export function parseEntryMolecules(json: unknown): string[] {
-  const entities = isRecord(json) ? Object.values(json)[0] : undefined
-  if (!Array.isArray(entities)) {
-    return []
-  }
-  return entities.flatMap(entity => {
-    if (!isRecord(entity)) {
-      return []
-    }
-    const { molecule_type: type, sequence } = entity
-    return typeof type === 'string' &&
-      type.startsWith('polypeptide') &&
-      typeof sequence === 'string' &&
-      sequence.length > 0
-      ? [sequence]
-      : []
-  })
-}

package/src/siftsNumbering.ts

@@ -1,61 +0,0 @@
-// How a PDB entry numbers the residues of a UniProt sequence. Author numbering
-// is what a paper cites and what a viewer lights a named residue by, and it is
-// not the UniProt numbering for every entry: haemoglobin's chains number from
-// the mature protein, so author residue 1 is UniProt residue 2 (2HHB, measured
-// 2026-09-12). SIFTS records the correspondence per chain segment, and
-// `fetchUniProtStructureMappings` is what asks PDBe for it.
-
-import type { UniProtStructureSegment } from './pdbUniProtMapping.ts'
-
-export interface AuthorRange {
-  start: number
-  end: number
-  chain: string
-  /** author number minus UniProt number in the chosen segment */
-  shift: number
-}
-
-/** The segments SIFTS gives one accession, chain ids and author numbers
- * included, out of a whole entry's mappings. */
-export function segmentsForAccession(
-  mappings: readonly {
-    accession: string
-    segments: UniProtStructureSegment[]
-  }[],
-  accession: string,
-) {
-  return mappings.find(m => m.accession === accession)?.segments ?? []
-}
-
-/**
- * The author-numbered range a UniProt range is cited by in the entry: the
- * segment overlapping most of the range decides the shift. Undefined when no
- * chain of the entry covers any of it, or when SIFTS named no author numbers.
- */
-export function toAuthorRange(
-  segments: readonly UniProtStructureSegment[],
-  range: { start: number; end: number },
-): AuthorRange | undefined {
-  const overlap = (s: UniProtStructureSegment) =>
-    Math.min(s.unpEnd, range.end) - Math.max(s.unpStart, range.start) + 1
-  const best = segments
-    .flatMap(s => {
-      const { chainId, authorStart, unpStart, unpEnd } = s
-      return chainId !== undefined &&
-        authorStart !== undefined &&
-        overlap(s) > 0
-        ? [{ chainId, authorStart, unpStart, unpEnd, overlap: overlap(s) }]
-        : []
-    })
-    .toSorted((a, b) => b.overlap - a.overlap)[0]
-  if (!best) {
-    return undefined
-  }
-  const shift = best.authorStart - best.unpStart
-  return {
-    start: Math.max(range.start, best.unpStart) + shift,
-    end: Math.min(range.end, best.unpEnd) + shift,
-    chain: best.chainId,
-    shift,
-  }
-}