@jbrowse/plugin-gff3 1.5.0 → 1.5.4

package/dist/plugin-gff3.esm.js

@@ -2,6 +2,7 @@ import AdapterType from '@jbrowse/core/pluggableElementTypes/AdapterType';
 import Plugin from '@jbrowse/core/Plugin';
 import { types } from 'mobx-state-tree';
 import { ConfigurationSchema, readConfObject } from '@jbrowse/core/configuration';
+import { getFileName, makeIndex, makeIndexType } from '@jbrowse/core/util/tracks';
 import { BaseFeatureDataAdapter } from '@jbrowse/core/data_adapters/BaseAdapter';
 import { doesIntersect2 } from '@jbrowse/core/util/range';
 import { openLocation } from '@jbrowse/core/util/io';
@@ -10,20 +11,16 @@ import SimpleFeature from '@jbrowse/core/util/simpleFeature';
 import { TabixIndexedFile } from '@gmod/tabix';
 import gff from '@gmod/gff';
 import IntervalTree from '@flatten-js/interval-tree';
+import { unzip } from '@gmod/bgzf-filehandle';
 
 function ownKeys(object, enumerableOnly) {
   var keys = Object.keys(object);
 
   if (Object.getOwnPropertySymbols) {
     var symbols = Object.getOwnPropertySymbols(object);
-
-    if (enumerableOnly) {
-      symbols = symbols.filter(function (sym) {
-        return Object.getOwnPropertyDescriptor(object, sym).enumerable;
-      });
-    }
-
-    keys.push.apply(keys, symbols);
+    enumerableOnly && (symbols = symbols.filter(function (sym) {
+      return Object.getOwnPropertyDescriptor(object, sym).enumerable;
+    })), keys.push.apply(keys, symbols);
   }
 
   return keys;
@@ -31,19 +28,12 @@ function ownKeys(object, enumerableOnly) {
 
 function _objectSpread2(target) {
   for (var i = 1; i < arguments.length; i++) {
-    var source = arguments[i] != null ? arguments[i] : {};
-
-    if (i % 2) {
-      ownKeys(Object(source), true).forEach(function (key) {
-        _defineProperty(target, key, source[key]);
-      });
-    } else if (Object.getOwnPropertyDescriptors) {
-      Object.defineProperties(target, Object.getOwnPropertyDescriptors(source));
-    } else {
-      ownKeys(Object(source)).forEach(function (key) {
-        Object.defineProperty(target, key, Object.getOwnPropertyDescriptor(source, key));
-      });
-    }
+    var source = null != arguments[i] ? arguments[i] : {};
+    i % 2 ? ownKeys(Object(source), !0).forEach(function (key) {
+      _defineProperty(target, key, source[key]);
+    }) : Object.getOwnPropertyDescriptors ? Object.defineProperties(target, Object.getOwnPropertyDescriptors(source)) : ownKeys(Object(source)).forEach(function (key) {
+      Object.defineProperty(target, key, Object.getOwnPropertyDescriptor(source, key));
+    });
   }
 
   return target;
@@ -104,6 +94,9 @@ function _defineProperties(target, props) {
 function _createClass(Constructor, protoProps, staticProps) {
   if (protoProps) _defineProperties(Constructor.prototype, protoProps);
   if (staticProps) _defineProperties(Constructor, staticProps);
+  Object.defineProperty(Constructor, "prototype", {
+    writable: false
+  });
   return Constructor;
 }
 
@@ -127,12 +120,15 @@ function _inherits(subClass, superClass) {
     throw new TypeError("Super expression must either be null or a function");
   }
 
-  subClass.prototype = Object.create(superClass && superClass.prototype, {
-    constructor: {
-      value: subClass,
-      writable: true,
-      configurable: true
-    }
+  Object.defineProperty(subClass, "prototype", {
+    value: Object.create(superClass && superClass.prototype, {
+      constructor: {
+        value: subClass,
+        writable: true,
+        configurable: true
+      }
+    }),
+    writable: false
   });
   if (superClass) _setPrototypeOf(subClass, superClass);
 }
@@ -333,6 +329,28 @@ var _default = /*#__PURE__*/function (_Plugin) {
           }
         });
       });
+      pluginManager.addToExtensionPoint('Core-guessAdapterForLocation', function (adapterGuesser) {
+        return function (file, index, adapterHint) {
+          var regexGuess = /\.gff3?\.b?gz$/i;
+          var adapterName = 'Gff3TabixAdapter';
+          var fileName = getFileName(file);
+          var indexName = index && getFileName(index);
+
+          if (regexGuess.test(fileName) || adapterHint === adapterName) {
+            return {
+              type: adapterName,
+              bamLocation: file,
+              gffGzLocation: file,
+              index: {
+                location: index || makeIndex(file, '.tbi'),
+                indexType: makeIndexType(indexName, 'CSI', 'TBI')
+              }
+            };
+          }
+
+          return adapterGuesser(file, index, adapterHint);
+        };
+      });
       pluginManager.addAdapterType(function () {
         return new AdapterType({
           name: 'Gff3Adapter',
@@ -344,6 +362,22 @@ var _default = /*#__PURE__*/function (_Plugin) {
           }
         });
       });
+      pluginManager.addToExtensionPoint('Core-guessAdapterForLocation', function (adapterGuesser) {
+        return function (file, index, adapterHint) {
+          var regexGuess = /\.gff3?$/i;
+          var adapterName = 'Gff3Adapter';
+          var fileName = getFileName(file);
+
+          if (regexGuess.test(fileName) || adapterHint === adapterName) {
+            return {
+              type: adapterName,
+              gffLocation: file
+            };
+          }
+
+          return adapterGuesser(file, index, adapterHint);
+        };
+      });
     }
   }]);
 
@@ -1255,9 +1289,6 @@ var _default$1 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
                   break;
                 }
 
-                // console.log(
-                //   `redispatching ${query.start}-${query.end} => ${minStart}-${maxEnd}`,
-                // )
                 // make a new feature callback to only return top-level features
                 // in the original query range
                 this.getFeaturesHelper(_objectSpread2(_objectSpread2({}, query), {}, {
@@ -1334,6 +1365,7 @@ var _default$1 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
       return featureLocs.map(function (featureLoc) {
         return new SimpleFeature({
           data: _this4.featureData(featureLoc),
+          // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
           id: "".concat(_this4.id, "-offset-").concat(featureLoc.attributes._lineHash[0])
         });
       });
@@ -1347,12 +1379,15 @@ var _default$1 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
 
       f.start -= 1; // convert to interbase
 
-      f.strand = {
-        '+': 1,
-        '-': -1,
-        '.': 0,
-        '?': undefined
-      }[data.strand]; // convert strand
+      if (data.strand === '+') {
+        f.strand = 1;
+      } else if (data.strand === '-') {
+        f.strand = -1;
+      } else if (data.strand === '.') {
+        f.strand = 0;
+      } else {
+        f.strand = undefined;
+      }
 
       f.phase = Number(data.phase);
       f.refName = data.seq_id;
@@ -1366,7 +1401,8 @@ var _default$1 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
       }
 
       var defaultFields = ['start', 'end', 'seq_id', 'score', 'type', 'source', 'phase', 'strand'];
-      Object.keys(data.attributes).forEach(function (a) {
+      var dataAttributes = data.attributes || {};
+      Object.keys(dataAttributes).forEach(function (a) {
         var b = a.toLowerCase();
 
         if (defaultFields.includes(b)) {
@@ -1375,8 +1411,8 @@ var _default$1 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
           b += '2';
         }
 
-        if (data.attributes[a] !== null) {
-          var attr = data.attributes[a];
+        if (dataAttributes[a] !== null) {
+          var attr = dataAttributes[a];
 
           if (Array.isArray(attr) && attr.length === 1) {
             var _attr = attr;
@@ -1409,7 +1445,9 @@ var _default$1 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
     }
   }, {
     key: "freeResources",
-    value: function freeResources() {}
+    value: function
+      /* { region } */
+    freeResources() {}
   }]);
 
   return _default;
@@ -1420,88 +1458,141 @@ var Gff3TabixAdapter = {
   'default': _default$1
 };
 
+function isGzip(buf) {
+  return buf[0] === 31 && buf[1] === 139 && buf[2] === 8;
+}
+
 var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
   _inherits(_default, _BaseFeatureDataAdapt);
 
   var _super = /*#__PURE__*/_createSuper(_default);
 
-  function _default(config, getSubAdapter, pluginManager) {
-    var _this;
-
+  function _default() {
     _classCallCheck(this, _default);
 
-    _this = _super.call(this, config, getSubAdapter, pluginManager);
-    var gffLocation = readConfObject(config, 'gffLocation');
-    var uri = gffLocation.uri;
-    _this.uri = uri;
-    _this.filehandle = openLocation(gffLocation, _this.pluginManager);
-    return _this;
+    return _super.apply(this, arguments);
   }
 
   _createClass(_default, [{
-    key: "loadData",
+    key: "loadDataP",
     value: function () {
-      var _loadData = /*#__PURE__*/_asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee() {
-        var _this2 = this;
-
-        var _yield$this$filehandl, size;
+      var _loadDataP = /*#__PURE__*/_asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee() {
+        var _this = this;
 
+        var buffer, buf, data, lines, headerLines, i, header, feats, intervalTree;
         return runtime_1.wrap(function _callee$(_context) {
           while (1) {
             switch (_context.prev = _context.next) {
               case 0:
                 _context.next = 2;
-                return this.filehandle.stat();
+                return openLocation(readConfObject(this.config, 'gffLocation'), this.pluginManager).readFile();
 
               case 2:
-                _yield$this$filehandl = _context.sent;
-                size = _yield$this$filehandl.size;
+                buffer = _context.sent;
 
-                if (!(size > 500000000)) {
-                  _context.next = 6;
+                if (!isGzip(buffer)) {
+                  _context.next = 9;
                   break;
                 }
 
-                throw new Error('This file is too large. Consider using Gff3TabixAdapter');
+                _context.next = 6;
+                return unzip(buffer);
 
               case 6:
+                _context.t0 = _context.sent;
+                _context.next = 10;
+                break;
+
+              case 9:
+                _context.t0 = buffer;
+
+              case 10:
+                buf = _context.t0;
+
+                if (!(buf.length > 536870888)) {
+                  _context.next = 13;
+                  break;
+                }
+
+                throw new Error('Data exceeds maximum string length (512MB)');
+
+              case 13:
+                data = new TextDecoder('utf8', {
+                  fatal: true
+                }).decode(buf);
+                lines = data.split('\n');
+                headerLines = [];
+
+                for (i = 0; i < lines.length && lines[i].startsWith('#'); i++) {
+                  headerLines.push(lines[i]);
+                }
+
+                header = headerLines.join('\n');
+                feats = gff.parseStringSync(data, {
+                  parseFeatures: true,
+                  parseComments: false,
+                  parseDirectives: false,
+                  parseSequences: false
+                });
+                intervalTree = feats.flat().map(function (f, i) {
+                  return new SimpleFeature({
+                    data: _this.featureData(f),
+                    id: "".concat(_this.id, "-offset-").concat(i)
+                  });
+                }).reduce(function (acc, obj) {
+                  var key = obj.get('refName');
+
+                  if (!acc[key]) {
+                    acc[key] = new IntervalTree();
+                  }
+
+                  acc[key].insert([obj.get('start'), obj.get('end')], obj);
+                  return acc;
+                }, {});
+                return _context.abrupt("return", {
+                  header: header,
+                  intervalTree: intervalTree
+                });
+
+              case 21:
+              case "end":
+                return _context.stop();
+            }
+          }
+        }, _callee, this);
+      }));
+
+      function loadDataP() {
+        return _loadDataP.apply(this, arguments);
+      }
+
+      return loadDataP;
+    }()
+  }, {
+    key: "loadData",
+    value: function () {
+      var _loadData = /*#__PURE__*/_asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee2() {
+        var _this2 = this;
+
+        return runtime_1.wrap(function _callee2$(_context2) {
+          while (1) {
+            switch (_context2.prev = _context2.next) {
+              case 0:
                 if (!this.gffFeatures) {
-                  this.gffFeatures = this.filehandle.readFile('utf8').then(function (data) {
-                    var gffFeatures = gff.parseStringSync(data, {
-                      parseFeatures: true,
-                      parseComments: false,
-                      parseDirectives: false,
-                      parseSequences: false
-                    });
-                    return gffFeatures.flat().map(function (f, i) {
-                      return new SimpleFeature({
-                        data: _this2.featureData(f),
-                        id: "".concat(_this2.id, "-offset-").concat(i)
-                      });
-                    }).reduce(function (acc, obj) {
-                      var key = obj.get('refName');
-
-                      if (!acc[key]) {
-                        acc[key] = new IntervalTree();
-                      }
-
-                      acc[key].insert([obj.get('start'), obj.get('end')], obj);
-                      return acc;
-                    }, {});
-                  })["catch"](function (e) {
+                  this.gffFeatures = this.loadDataP()["catch"](function (e) {
                     _this2.gffFeatures = undefined;
                     throw e;
                   });
                 }
 
-                return _context.abrupt("return", this.gffFeatures);
+                return _context2.abrupt("return", this.gffFeatures);
 
-              case 8:
+              case 2:
               case "end":
-                return _context.stop();
+                return _context2.stop();
             }
           }
-        }, _callee, this);
+        }, _callee2, this);
       }));
 
       function loadData() {
@@ -1513,25 +1604,28 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
   }, {
     key: "getRefNames",
     value: function () {
-      var _getRefNames = /*#__PURE__*/_asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee2() {
-        var gffFeatures;
-        return runtime_1.wrap(function _callee2$(_context2) {
+      var _getRefNames = /*#__PURE__*/_asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee3() {
+        var _yield$this$loadData,
+            intervalTree;
+
+        return runtime_1.wrap(function _callee3$(_context3) {
           while (1) {
-            switch (_context2.prev = _context2.next) {
+            switch (_context3.prev = _context3.next) {
               case 0:
-                _context2.next = 3;
+                _context3.next = 3;
                 return this.loadData();
 
               case 3:
-                gffFeatures = _context2.sent;
-                return _context2.abrupt("return", Object.keys(gffFeatures));
+                _yield$this$loadData = _context3.sent;
+                intervalTree = _yield$this$loadData.intervalTree;
+                return _context3.abrupt("return", Object.keys(intervalTree));
 
-              case 5:
+              case 6:
               case "end":
-                return _context2.stop();
+                return _context3.stop();
             }
           }
-        }, _callee2, this);
+        }, _callee3, this);
       }));
 
       function getRefNames() {
@@ -1540,6 +1634,38 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
 
       return getRefNames;
     }()
+  }, {
+    key: "getHeader",
+    value: function () {
+      var _getHeader = /*#__PURE__*/_asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee4() {
+        var _yield$this$loadData2, header;
+
+        return runtime_1.wrap(function _callee4$(_context4) {
+          while (1) {
+            switch (_context4.prev = _context4.next) {
+              case 0:
+                _context4.next = 2;
+                return this.loadData();
+
+              case 2:
+                _yield$this$loadData2 = _context4.sent;
+                header = _yield$this$loadData2.header;
+                return _context4.abrupt("return", header);
+
+              case 5:
+              case "end":
+                return _context4.stop();
+            }
+          }
+        }, _callee4, this);
+      }));
+
+      function getHeader() {
+        return _getHeader.apply(this, arguments);
+      }
+
+      return getHeader;
+    }()
   }, {
     key: "getFeatures",
     value: function getFeatures(query) {
@@ -1547,39 +1673,39 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
 
       var opts = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
       return ObservableCreate( /*#__PURE__*/function () {
-        var _ref = _asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee3(observer) {
-          var start, end, refName, gffFeatures, tree, feats;
-          return runtime_1.wrap(function _callee3$(_context3) {
+        var _ref = _asyncToGenerator( /*#__PURE__*/runtime_1.mark(function _callee5(observer) {
+          var _intervalTree$refName, start, end, refName, _yield$_this3$loadDat, intervalTree;
+
+          return runtime_1.wrap(function _callee5$(_context5) {
             while (1) {
-              switch (_context3.prev = _context3.next) {
+              switch (_context5.prev = _context5.next) {
                 case 0:
-                  _context3.prev = 0;
+                  _context5.prev = 0;
                   start = query.start, end = query.end, refName = query.refName;
-                  _context3.next = 4;
+                  _context5.next = 4;
                   return _this3.loadData();
 
                 case 4:
-                  gffFeatures = _context3.sent;
-                  tree = gffFeatures[refName];
-                  feats = tree.search([start, end]);
-                  feats.forEach(function (f) {
-                    observer.next(f);
+                  _yield$_this3$loadDat = _context5.sent;
+                  intervalTree = _yield$_this3$loadDat.intervalTree;
+                  (_intervalTree$refName = intervalTree[refName]) === null || _intervalTree$refName === void 0 ? void 0 : _intervalTree$refName.search([start, end]).forEach(function (f) {
+                    return observer.next(f);
                   });
                   observer.complete();
-                  _context3.next = 14;
+                  _context5.next = 13;
                   break;
 
-                case 11:
-                  _context3.prev = 11;
-                  _context3.t0 = _context3["catch"](0);
-                  observer.error(_context3.t0);
+                case 10:
+                  _context5.prev = 10;
+                  _context5.t0 = _context5["catch"](0);
+                  observer.error(_context5.t0);
 
-                case 14:
+                case 13:
                 case "end":
-                  return _context3.stop();
+                  return _context5.stop();
               }
             }
-          }, _callee3, null, [[0, 11]]);
+          }, _callee5, null, [[0, 10]]);
         }));
 
         return function (_x) {
@@ -1596,12 +1722,15 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
 
       f.start -= 1; // convert to interbase
 
-      f.strand = {
-        '+': 1,
-        '-': -1,
-        '.': 0,
-        '?': undefined
-      }[data.strand]; // convert strand
+      if (data.strand === '+') {
+        f.strand = 1;
+      } else if (data.strand === '-') {
+        f.strand = -1;
+      } else if (data.strand === '.') {
+        f.strand = 0;
+      } else {
+        f.strand = undefined;
+      }
 
       f.phase = Number(data.phase);
       f.refName = data.seq_id;
@@ -1615,7 +1744,8 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
       }
 
       var defaultFields = ['start', 'end', 'seq_id', 'score', 'type', 'source', 'phase', 'strand'];
-      Object.keys(data.attributes).forEach(function (a) {
+      var dataAttributes = data.attributes || {};
+      Object.keys(dataAttributes).forEach(function (a) {
         var b = a.toLowerCase();
 
         if (defaultFields.includes(b)) {
@@ -1624,8 +1754,8 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
           b += '2';
         }
 
-        if (data.attributes[a] !== null) {
-          var attr = data.attributes[a];
+        if (dataAttributes[a] !== null) {
+          var attr = dataAttributes[a];
 
           if (Array.isArray(attr) && attr.length === 1) {
             var _attr = attr;
@@ -1657,7 +1787,9 @@ var _default$2 = /*#__PURE__*/function (_BaseFeatureDataAdapt) {
     }
   }, {
     key: "freeResources",
-    value: function freeResources() {}
+    value: function
+      /* { region } */
+    freeResources() {}
   }]);
 
   return _default;