-
Notifications
You must be signed in to change notification settings - Fork 128
/
ChunkedFileData.js
217 lines (188 loc) · 6.49 KB
/
ChunkedFileData.js
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
/**
* This class represents a file that might not have all its data loaded yet.
* It is used when loading the entire file is not an option because it's too
* expensive. Instead, parts of the file are loaded and added only when needed.
* From a reading point of view is as if the entire file is loaded. The
* exception is when the data is not available yet, an error will be thrown.
* This class does not load the data, it just manages it. It provides operations
* to add and read data from the file.
*
* @flow
*/
'use strict';
const NOT_FOUND = -1;
import type {
ChunkType,
DataType
} from './FlowTypes';
class ChunkedFileData {
// $FlowIssue - get/set properties not yet supported
static get NOT_FOUND() { return NOT_FOUND; }
_fileData: Array<ChunkType>;
constructor() {
this._fileData = [];
}
/**
* Adds data to the file storage at a specific offset.
*/
addData(offset: number, data: DataType): void {
var offsetEnd = offset+data.length-1;
var chunkRange = this._getChunkRange(offset, offsetEnd);
if (chunkRange.startIx === NOT_FOUND) {
this._fileData.splice(chunkRange.insertIx || 0, 0, {
offset: offset,
data: data
});
} else {
// If the data to add collides with existing chunks we prepend and
// append data from the half colliding chunks to make the collision at
// 100%. The new data can then replace all the colliding chunkes.
var firstChunk = this._fileData[chunkRange.startIx];
var lastChunk = this._fileData[chunkRange.endIx];
var needsPrepend = offset > firstChunk.offset;
var needsAppend = offsetEnd < lastChunk.offset + lastChunk.data.length - 1;
var chunk = {
offset: Math.min(offset, firstChunk.offset),
data: data
};
if (needsPrepend) {
var slicedData = this._sliceData(
firstChunk.data,
0,
offset - firstChunk.offset
);
chunk.data = this._concatData(slicedData, data);
}
if (needsAppend) {
// Use the lastChunk because the slice logic is easier to handle.
var slicedData = this._sliceData(
chunk.data,
0,
lastChunk.offset - chunk.offset
);
chunk.data = this._concatData(slicedData, lastChunk.data);
}
this._fileData.splice(
chunkRange.startIx,
chunkRange.endIx - chunkRange.startIx + 1,
chunk
);
}
}
_concatData(dataA: DataType, dataB: DataType): DataType {
// TypedArrays don't support concat.
if (
typeof ArrayBuffer !== "undefined" && ArrayBuffer.isView &&
ArrayBuffer.isView(dataA)
) {
// $FlowIssue - flow thinks dataAandB is a string but it's not
var dataAandB = new dataA.constructor(dataA.length + dataB.length);
// $FlowIssue - flow thinks dataAandB is a string but it's not
dataAandB.set(dataA, 0);
// $FlowIssue - flow thinks dataAandB is a string but it's not
dataAandB.set(dataB, dataA.length);
return dataAandB;
} else {
// $FlowIssue - flow thinks dataAandB is a TypedArray but it's not
return dataA.concat(dataB);
}
}
_sliceData(data: DataType, begin: number, end: number): DataType {
// Some TypeArray implementations do not support slice yet.
if (data.slice) {
return data.slice(begin, end);
} else {
// $FlowIssue - flow thinks data is a string but it's not
return data.subarray(begin, end);
}
}
/**
* Finds the chunk range that overlaps the [offsetStart-1,offsetEnd+1] range.
* When a chunk is adjacent to the offset we still consider it part of the
* range (this is the situation of offsetStart-1 or offsetEnd+1).
* When no chunks are found `insertIx` denotes the index where the data
* should be inserted in the data list (startIx == NOT_FOUND and endIX ==
* NOT_FOUND).
*/
_getChunkRange(
offsetStart: number,
offsetEnd: number
): {startIx: number, endIx: number, insertIx?: number} {
var startChunkIx = NOT_FOUND;
var endChunkIx = NOT_FOUND;
var insertIx = 0;
// Could use binary search but not expecting that many blocks to exist.
for (var i = 0; i < this._fileData.length; i++, insertIx = i) {
var chunkOffsetStart = this._fileData[i].offset;
var chunkOffsetEnd = chunkOffsetStart + this._fileData[i].data.length;
if (offsetEnd < chunkOffsetStart-1) {
// This offset range doesn't overlap with any chunks.
break;
}
// If it is adjacent we still consider it part of the range because
// we're going end up with a single block with all contiguous data.
if (offsetStart <= chunkOffsetEnd+1 &&
offsetEnd >= chunkOffsetStart-1) {
startChunkIx = i;
break;
}
}
// No starting chunk was found, meaning that the offset is either before
// or after the current stored chunks.
if (startChunkIx === NOT_FOUND) {
return {
startIx: NOT_FOUND,
endIx: NOT_FOUND,
insertIx: insertIx
};
}
// Find the ending chunk.
for (var i = startChunkIx; i < this._fileData.length; i++) {
var chunkOffsetStart = this._fileData[i].offset;
var chunkOffsetEnd = chunkOffsetStart + this._fileData[i].data.length;
if (offsetEnd >= chunkOffsetStart-1) {
// Candidate for the end chunk, it doesn't mean it is yet.
endChunkIx = i;
}
if (offsetEnd <= chunkOffsetEnd+1) {
break;
}
}
if (endChunkIx === NOT_FOUND) {
endChunkIx = startChunkIx;
}
return {
startIx: startChunkIx,
endIx: endChunkIx
};
}
hasDataRange(offsetStart: number, offsetEnd: number): boolean {
for (var i = 0; i < this._fileData.length; i++) {
var chunk = this._fileData[i];
if (offsetEnd < chunk.offset) {
return false;
}
if (offsetStart >= chunk.offset &&
offsetEnd < chunk.offset + chunk.data.length) {
return true;
}
}
return false;
}
getByteAt(offset: number): any {
var dataChunk;
for (var i = 0; i < this._fileData.length; i++) {
var dataChunkStart = this._fileData[i].offset;
var dataChunkEnd = dataChunkStart + this._fileData[i].data.length - 1;
if (offset >= dataChunkStart && offset <= dataChunkEnd) {
dataChunk = this._fileData[i];
break;
}
}
if (dataChunk) {
return dataChunk.data[offset - dataChunk.offset];
}
throw new Error("Offset " + offset + " hasn't been loaded yet.");
}
}
module.exports = ChunkedFileData;