@datagrok/bio 2.4.50 → 2.4.52

package/package.json

@@ -5,7 +5,7 @@
     "name": "Leonid Stolbov",
     "email": "lstolbov@datagrok.ai"
   },
-  "version": "2.4.50",
+  "version": "2.4.52",
   "description": "Bioinformatics support (import/export of sequences, conversion, visualization, analysis). [See more](https://github.com/datagrok-ai/public/blob/master/packages/Bio/README.md) for details.",
   "repository": {
     "type": "git",
@@ -22,7 +22,7 @@
   ],
   "dependencies": {
     "@biowasm/aioli": "^3.1.0",
-    "@datagrok-libraries/bio": "^5.32.7",
+    "@datagrok-libraries/bio": "^5.32.8",
     "@datagrok-libraries/chem-meta": "^1.0.1",
     "@datagrok-libraries/ml": "^6.3.39",
     "@datagrok-libraries/tutorials": "^1.3.2",

package/src/package.ts

@@ -614,6 +614,16 @@ export function importFasta(fileContent: string): DG.DataFrame [] {
   const ffh = new FastaFileHandler(fileContent);
   return ffh.importFasta();
 }
+//name: importBam
+//description: Opens Bam file
+//tags: file-handler
+//meta.ext: bam, bai
+//input: string fileContent
+//output: list tables
+export function importBam(fileContent: string): DG.DataFrame [] {
+  console.log(fileContent);
+  return [];
+}
 
 //top-menu: Bio | Convert...
 //name: convertDialog

package/src/tests/detectors-benchmark-tests.ts

@@ -39,17 +39,15 @@ category('detectorsBenchmark', () => {
 
   test('separatorDnaShorts50Few50', async () => {
     detectMacromoleculeBenchmark(10, NOTATION.SEPARATOR, ALPHABET.DNA, 50, 50, '/');
-  }, {skipReason: '#1192'});
+  });
 
   test('separatorDnaShorts50Many1E6', async () => {
-      detectMacromoleculeBenchmark(10, NOTATION.SEPARATOR, ALPHABET.DNA, 50, 1E6, '/');
-    },
-    { /* skipReason: 'slow transmit large dataset to detector' */});
+    detectMacromoleculeBenchmark(10, NOTATION.SEPARATOR, ALPHABET.DNA, 50, 1E6, '/');
+  });
 
   test('separatorDnaLong1e6Few50', async () => {
-      detectMacromoleculeBenchmark(10, NOTATION.SEPARATOR, ALPHABET.DNA, 1E6, 50, '/');
-    },
-    {skipReason: '#1192'});
+    detectMacromoleculeBenchmark(10, NOTATION.SEPARATOR, ALPHABET.DNA, 1E6, 50, '/');
+  });
 
   async function detectMacromoleculeBenchmark(
     maxET: number, notation: NOTATION, alphabet: ALPHABET, length: number, count: number, separator?: string,
@@ -128,10 +126,10 @@ category('detectorsBenchmark', () => {
 
   function checkDetectorRes(col: DG.Column, tgt: TgtType): void {
     const uh = UnitsHandler.getOrCreate(col);
-    expect(col.semType, tgt.semType);
-    expect(uh.notation, tgt.notation);
-    expect(uh.alphabet, tgt.alphabet);
-    expect(uh.separator, tgt.separator);
+    expect(col.semType === tgt.semType, true);
+    expect(uh.notation === tgt.notation, true);
+    expect(uh.alphabet === tgt.alphabet, true);
+    expect(uh.separator === tgt.separator, true);
   }
 });
 

package/src/tests/renderers-monomer-placer.ts

@@ -3,7 +3,7 @@ import * as DG from 'datagrok-api/dg';
 import * as ui from 'datagrok-api/ui';
 
 import {_package} from '../package-test';
-import {after, before, category, delay, expect, test} from '@datagrok-libraries/utils/src/test';
+import {category, test} from '@datagrok-libraries/utils/src/test';
 import {MonomerPlacer} from '@datagrok-libraries/bio/src/utils/cell-renderer-monomer-placer';
 import {monomerToShort} from '@datagrok-libraries/bio/src/utils/macromolecule';
 import {UnitsHandler} from '@datagrok-libraries/bio/src/utils/units-handler';
@@ -22,13 +22,14 @@ category('renderers: monomerPlacer', () => {
         'id3,mon1-M-mon3-mon4-mon5-MON8-N9\n', //
       testList: [
         {src: {row: 0, x: -1}, tgt: {pos: null}},
-        {src: {row: 1, x: 0}, tgt: {pos: 0}},
-        {src: {row: 1, x: 1}, tgt: {pos: 0}},
+        {src: {row: 1, x: 0}, tgt: {pos: null}},
+        {src: {row: 1, x: 5}, tgt: {pos: 0}},
+        {src: {row: 1, x: 6}, tgt: {pos: 0}},
         {src: {row: 1, x: 26}, tgt: {pos: 1}},
         {src: {row: 1, x: 170}, tgt: {pos: 6}},
-        {src: {row: 1, x: 208}, tgt: {pos: null}},
-        {src: {row: 2, x: 170}, tgt: {pos: 6}},
-        {src: {row: 2, x: 175}, tgt: {pos: null}},
+        {src: {row: 1, x: 208}, tgt: {pos: 7}},
+        {src: {row: 2, x: 170}, tgt: {pos: 5}},
+        {src: {row: 2, x: 175}, tgt: {pos: 5}},
       ]
     },
     splitterMsa: {
@@ -41,13 +42,13 @@ category('renderers: monomerPlacer', () => {
         'id3,mon1-M-mon3-mon4-mon5---MON8-N9\n', //
       testList: [
         {src: {row: 0, x: -1}, tgt: {pos: null}},
-        {src: {row: 1, x: 0}, tgt: {pos: 0}},
-        {src: {row: 1, x: 1}, tgt: {pos: 0}},
-        {src: {row: 1, x: 26}, tgt: {pos: 1}},
-        {src: {row: 1, x: 170}, tgt: {pos: 6}},
-        {src: {row: 1, x: 227}, tgt: {pos: null}},
-        {src: {row: 2, x: 220}, tgt: {pos: 8}},
-        {src: {row: 2, x: 227}, tgt: {pos: null}},
+        {src: {row: 1, x: 0}, tgt: {pos: null}},
+        {src: {row: 1, x: 1}, tgt: {pos: null}},
+        {src: {row: 1, x: 26}, tgt: {pos: 0}},
+        {src: {row: 1, x: 170}, tgt: {pos: 4}},
+        {src: {row: 1, x: 227}, tgt: {pos: 5}},
+        {src: {row: 2, x: 220}, tgt: {pos: 5}},
+        {src: {row: 2, x: 227}, tgt: {pos: 5}},
       ]
     },
     fastaMsa: {
@@ -61,13 +62,13 @@ id3,QHIRE--LT
 `,
       testList: [
         {src: {row: 1, x: -1}, tgt: {pos: null}},
-        {src: {row: 1, x: 0}, tgt: {pos: 0}},
-        {src: {row: 1, x: 1}, tgt: {pos: 0}},
-        {src: {row: 1, x: 19}, tgt: {pos: 1}},
+        {src: {row: 1, x: 0}, tgt: {pos: null}},
+        {src: {row: 1, x: 1}, tgt: {pos: null}},
+        {src: {row: 1, x: 19}, tgt: {pos: 0}},
         {src: {row: 1, x: 170}, tgt: {pos: 8}},
-        {src: {row: 1, x: 171}, tgt: {pos: null}},
+        {src: {row: 1, x: 171}, tgt: {pos: 8}},
         {src: {row: 2, x: 170}, tgt: {pos: 8}},
-        {src: {row: 2, x: 171}, tgt: {pos: null}},
+        {src: {row: 2, x: 181}, tgt: {pos: null}},
       ]
     },
   };
@@ -78,7 +79,7 @@ id3,QHIRE--LT
       await grok.data.detectSemanticTypes(df);
       const seqCol: DG.Column = df.getCol('seq');
 
-      const monLength: number = 1;
+      const monLength: number = 3;
       const charWidth: number = 7;
       const sepWidth: number = 12;
       const colTemp: MonomerPlacer = new MonomerPlacer(null, seqCol, () => {

package/src/utils/cell-renderer-consts.ts

@@ -8,15 +8,28 @@ export enum MonomerWidthMode {
 }
 
 export const enum Tags {
-  calculated = '.mm.cellRenderer.calculated',
+  RendererSettingsChanged = '.mm.cellRenderer.settingsChanged',
 }
 
+export const rendererSettingsChangedState = {
+  true: '1',
+  false: '0',
+};
+
 export const enum Temps {
   monomerWidth = '.mm.cellRenderer.monomerWidth',
   maxMonomerLength = '.mm.cellRenderer.maxMonomerLength',
   colorCode = '.mm.cellRenderer.colorCode',
   compareWithCurrent = '.mm.cellRenderer.compareWithCurrent',
   highlightDifference = '.mm.cellRenderer.highlightDifference',
+  gapLength = '.mm.cellRenderer.gapLength',
+}
+
+export const enum tempTAGS {
+  referenceSequence = 'reference-sequence',
+  currentWord = 'current-word',
+  monomerWidth = 'monomer-width',
+  bioSeqCol = 'bio-seqCol',
 }
 
 // export const MacromoleculeCellRendererDefaults = new class {

package/src/utils/cell-renderer.ts

@@ -7,10 +7,8 @@ import {printLeftOrCentered, DrawStyle} from '@datagrok-libraries/bio/src/utils/
 import * as C from './constants';
 import {MonomerPlacer} from '@datagrok-libraries/bio/src/utils/cell-renderer-monomer-placer';
 import {
-  ALIGNMENT, ALPHABET,
   getPaletteByType,
   getSplitter,
-  getSplitterForColumn,
   monomerToShort,
   MonomerToShortFunc,
   NOTATION,
@@ -20,22 +18,8 @@ import {
 import {SeqPalette} from '@datagrok-libraries/bio/src/seq-palettes';
 import {UnknownSeqPalettes} from '@datagrok-libraries/bio/src/unknown';
 import {UnitsHandler} from '@datagrok-libraries/bio/src/utils/units-handler';
-import {MonomerWorks} from '@datagrok-libraries/bio/src/monomer-works/monomer-works';
-import {Tags as mmcrTags, Temps as mmcrTemps} from '../utils/cell-renderer-consts';
-import { HELM_POLYMER_TYPE } from '@datagrok-libraries/bio/src/utils/const';
-import { MonomerLib } from './monomer-lib';
-import { IMonomerLib } from '@datagrok-libraries/bio/src/types';
-
-const enum tempTAGS {
-  referenceSequence = 'reference-sequence',
-  currentWord = 'current-word',
-  monomerWidth = 'monomer-width',
-  bioSeqCol = 'bio-seqCol',
-}
-
-const enum rndrTAGS {
-  calculatedCellRender = '.calculatedCellRender',
-}
+import {Temps as mmcrTemps, Tags as mmcrTags,
+  tempTAGS, rendererSettingsChangedState} from '../utils/cell-renderer-consts';
 
 type TempType = { [tagName: string]: any };
 
@@ -85,7 +69,7 @@ export class MacromoleculeSequenceCellRenderer extends DG.GridCellRenderer {
     //   return;
 
     const tableCol: DG.Column = gridCell.cell.column;
-    const tableColTemp: TempType = tableCol.temp;
+    //const tableColTemp: TempType = tableCol.temp;
     const seqColTemp: MonomerPlacer = tableCol.temp[tempTAGS.bioSeqCol];
     if (!seqColTemp) return; // Can do nothing without precalculated data
 
@@ -105,10 +89,10 @@ export class MacromoleculeSequenceCellRenderer extends DG.GridCellRenderer {
       const monomerSymbol: string = seqMonList[left];
       const tooltipElements: HTMLElement[] = [ui.div(monomerSymbol)];
       const monomer = seqColTemp.getMonomer(monomerSymbol);
-      if(monomer) {
-      const options = {autoCrop: true, autoCropMargin: 0, suppressChiralText: true};
-      const monomerSVG = grok.chem.svgMol(monomer.smiles, undefined, undefined, options);
-      tooltipElements.push(monomerSVG);
+      if (monomer) {
+        const options = {autoCrop: true, autoCropMargin: 0, suppressChiralText: true};
+        const monomerSVG = grok.chem.svgMol(monomer.smiles, undefined, undefined, options);
+        tooltipElements.push(monomerSVG);
       }
       ui.tooltip.show(ui.divV(tooltipElements), e.x + 16, e.y + 16);
     } else {
@@ -130,9 +114,10 @@ export class MacromoleculeSequenceCellRenderer extends DG.GridCellRenderer {
    */
   render(
     g: CanvasRenderingContext2D, x: number, y: number, w: number, h: number, gridCell: DG.GridCell,
-    cellStyle: DG.GridCellStyle
+    _cellStyle: DG.GridCellStyle
   ): void {
-    let gapRenderer = 5;
+    let gapLength = 0;
+    const msaGapLength = 8;
     let maxLengthOfMonomer = 8;
 
     // TODO: Store temp data to GridColumn
@@ -144,9 +129,9 @@ export class MacromoleculeSequenceCellRenderer extends DG.GridCellRenderer {
     // Cell renderer settings
     const tempMonomerWidth: string | null = tableColTemp[tempTAGS.monomerWidth];
     const monomerWidth: string = (tempMonomerWidth != null) ? tempMonomerWidth : 'short';
-    if (monomerWidth === 'short') {
+    if (monomerWidth === 'short')
       maxLengthOfMonomer = tableColTemp[mmcrTemps.maxMonomerLength] ?? _package.properties.maxMonomerLength;
-    }
+
 
     let seqColTemp: MonomerPlacer = tableCol.temp[tempTAGS.bioSeqCol];
     if (!seqColTemp) {
@@ -155,16 +140,24 @@ export class MacromoleculeSequenceCellRenderer extends DG.GridCellRenderer {
           const uh = UnitsHandler.getOrCreate(tableCol);
           return {
             unitsHandler: uh,
-            monomerCharWidth: 7, separatorWidth: !uh.isMsa() ? gapRenderer : 8,
+            monomerCharWidth: 7, separatorWidth: !uh.isMsa() ? gapLength : msaGapLength,
             monomerToShort: monomerToShortFunction, monomerLengthLimit: maxLengthOfMonomer,
             monomerLib: getBioLib()
           };
         });
     }
 
+    if (tableCol.tags[mmcrTags.RendererSettingsChanged] === rendererSettingsChangedState.true) {
+      gapLength = tableColTemp[mmcrTemps.gapLength] as number ?? gapLength;
+      // this event means that the mm renderer settings have changed, particularly monomer representation and max width.
+      seqColTemp.setMonomerLengthLimit(maxLengthOfMonomer);
+      seqColTemp.setSeparatorWidth(seqColTemp.isMsa() ? msaGapLength : gapLength);
+      tableCol.setTag(mmcrTags.RendererSettingsChanged, rendererSettingsChangedState.false);
+    }
+
     const [maxLengthWords, maxLengthWordsSum]: [number[], number[]] =
       seqColTemp.getCellMonomerLengths(gridCell.tableRowIndex!);
-    const maxIndex = maxLengthWords.length;
+    const _maxIndex = maxLengthWords.length;
 
     // Store updated seqColTemp to the col temp
     if (seqColTemp.updated) tableColTemp[tempTAGS.bioSeqCol] = seqColTemp;

package/src/widgets/representations.ts

@@ -2,7 +2,8 @@ import * as grok from 'datagrok-api/grok';
 import * as ui from 'datagrok-api/ui';
 import * as DG from 'datagrok-api/dg';
 import {getMolfilesFromSingleSeq} from '@datagrok-libraries/bio/src/monomer-works/monomer-utils';
-import {Tags as mmcrTags, Temps as mmcrTemps, MonomerWidthMode} from '../utils/cell-renderer-consts';
+import {Tags as mmcrTags, Temps as mmcrTemps, MonomerWidthMode,
+  tempTAGS, rendererSettingsChangedState} from '../utils/cell-renderer-consts';
 import {_package} from '../package';
 
 
@@ -19,11 +20,11 @@ export function getMacromoleculeColumnPropertyPanel(col: DG.Column): DG.Widget {
   columnsSet.delete(col.name);
 
   const monomerWidth = ui.choiceInput('Monomer width',
-    (col?.temp['monomer-width'] != null) ? col.temp['monomer-width'] : 'short',
-    ['short', 'long'],
+    (col?.temp[tempTAGS.monomerWidth] != null) ? col.temp[tempTAGS.monomerWidth] : MonomerWidthMode.short,
+    [MonomerWidthMode.short, MonomerWidthMode.long],
     (s: string) => {
-      col.temp['monomer-width'] = s;
-      col.setTag(mmcrTags.calculated, '0');
+      col.temp[tempTAGS.monomerWidth] = s;
+      col.setTag(mmcrTags.RendererSettingsChanged, rendererSettingsChangedState.true);
       col.dataFrame.fireValuesChanged();
     });
   monomerWidth.setTooltip(
@@ -34,13 +35,21 @@ export function getMacromoleculeColumnPropertyPanel(col: DG.Column): DG.Widget {
     col.temp[mmcrTemps.maxMonomerLength] ?? _package.properties.maxMonomerLength,
     (value: number) => {
       col.temp[mmcrTemps.maxMonomerLength] = value;
-      col.setTag(mmcrTags.calculated, '0');
+      col.setTag(mmcrTags.RendererSettingsChanged, rendererSettingsChangedState.true);
       col.dataFrame.fireValuesChanged();
     });
   maxMonomerLength.setTooltip(
     `The max length of monomer name displayed without shortening in '${MonomerWidthMode.short}' monomer width mode.`
   );
 
+  const gapLengthInput = ui.intInput('Monomer gap size', col.temp[mmcrTemps.gapLength] ?? 0,
+    (value: number) => {
+      col.temp[mmcrTemps.gapLength] = value;
+      col.setTag(mmcrTags.RendererSettingsChanged, rendererSettingsChangedState.true);
+      col.dataFrame.fireValuesChanged();
+    });
+  gapLengthInput.setTooltip('The size of Gap between monomers (in pixels)');
+
   const colorCode = ui.boolInput('Color code',
     (col?.temp['color-code'] != null) ? col.temp['color-code'] : true,
     (v: boolean) => {
@@ -68,8 +77,9 @@ export function getMacromoleculeColumnPropertyPanel(col: DG.Column): DG.Widget {
   const rdKitInputs = ui.inputs([
     monomerWidth,
     maxMonomerLength,
-    colorCode,
+    gapLengthInput,
     referenceSequence,
+    colorCode,
     compareWithCurrent,
   ]);