## Public and Private API ## _API documentation automatically generated by [docmeteor](https://github.com/raix/docmeteor)._ *** __File: ["tempStore.js"](tempStore.js) Where: {server}__ *** ##Temporary Storage Temporary storage is used for chunked uploads until all chunks are received and all copies have been made or given up. In some cases, the original file is stored only in temporary storage (for example, if all copies do some manipulation in beforeSave). This is why we use the temporary file as the basis for each saved copy, and then remove it after all copies are saved. Every chunk is saved as an individual temporary file. This is safer than attempting to write multiple incoming chunks to different positions in a single temporary file, which can lead to write conflicts. Using temp files also allows us to easily resume uploads, even if the server restarts, and to keep the working memory clear. The FS.TempStore emits events that others are able to listen to - ### *fs*.TempStore {object}  Server ### *This property __TempStore__ is defined in `FS`* it's an event emitter* > ```FS.TempStore = new EventEmitter();``` [tempStore.js:28](tempStore.js#L28) - ### *fsTempstore*.Storage {StorageAdapter}  Server ### *This property is private* *This property __Storage__ is defined in `FS.TempStore`* This property is set to either `FS.Store.FileSystem` or `FS.Store.GridFS` __When and why:__ We normally default to `cfs-filesystem` unless its not installed. *(we default to gridfs if installed)* But if `cfs-gridfs` and `cfs-worker` is installed we default to `cfs-gridfs` If `cfs-gridfs` and `cfs-filesystem` is not installed we log a warning. the user can set `FS.TempStore.Storage` them selfs eg.: ```js // Its important to set `internal: true` this lets the SA know that we // are using this internally and it will give us direct SA api FS.TempStore.Storage = new FS.Store.GridFS('_tempstore', { internal: true }); ``` > Note: This is considered as `advanced` use, its not a common pattern. > ```FS.TempStore.Storage = null;``` [tempStore.js:54](tempStore.js#L54) - We will not mount a storage adapter until needed. This allows us to check for the existance of FS.FileWorker, which is loaded after this package because it depends on this package. - XXX: TODO FS.TempStore.on('stored', function(fileObj, chunkCount, result) { This should work if we pass on result from the SA on stored event... fileObj.update({ $set: { chunkSum: 1, chunkCount: chunkCount, size: result.size } }); }); Stream implementation - ### _chunkPath([n])  Server ### *This method is private* __Arguments__ * __n__ *{Number}* (Optional) Chunk number __Returns__ *{String}* Chunk naming convention > ```_chunkPath = function(n) { ...``` [tempStore.js:104](tempStore.js#L104) - ### _fileReference(fileObj, chunk)  Server ### *This method is private* __Arguments__ * __fileObj__ *{[FS.File](#FS.File)}* * __chunk__ *{Number}* __Returns__ *{String}* Generated SA specific fileKey for the chunk Note: Calling function should call mountStorage() first, and make sure that fileObj is mounted. > ```_fileReference = function(fileObj, chunk, existing) { ...``` [tempStore.js:118](tempStore.js#L118) - ### *fsTempstore*.exists(File)  Server ### *This method __exists__ is defined in `FS.TempStore`* __Arguments__ * __File__ *{[FS.File](#FS.File)}* object __Returns__ *{Boolean}* Is this file, or parts of it, currently stored in the TempStore > ```FS.TempStore.exists = function(fileObj) { ...``` [tempStore.js:145](tempStore.js#L145) - ### *fsTempstore*.listParts(fileObj)  Server ### *This method __listParts__ is defined in `FS.TempStore`* __Arguments__ * __fileObj__ *{[FS.File](#FS.File)}* __Returns__ *{Object}* of parts already stored __TODO__ ``` * This is not yet implemented, milestone 1.1.0 ``` > ```FS.TempStore.listParts = function(fileObj) { ...``` [tempStore.js:156](tempStore.js#L156) - ### *fsTempstore*.removeFile(fileObj)  Server ### *This method __removeFile__ is defined in `FS.TempStore`* __Arguments__ * __fileObj__ *{[FS.File](#FS.File)}* This function removes the file from tempstorage - it cares not if file is already removed or not found, goal is reached anyway. > ```FS.TempStore.removeFile = function(fileObj) { ...``` [tempStore.js:169](tempStore.js#L169) - ### *fsTempstore*.createWriteStream(fileObj, [options])  Server ### *This method __createWriteStream__ is defined in `FS.TempStore`* __Arguments__ * __fileObj__ *{[FS.File](#FS.File)}* File to store in temporary storage * __options__ *{[Number ](#Number )|[ String](# String)}* (Optional) __Returns__ *{Stream}* Writeable stream `options` of different types mean differnt things: `undefined` We store the file in one part (Normal server-side api usage)* `Number` the number is the part number total (multipart uploads will use this api)* `String` the string is the name of the `store` that wants to store file data (stores that want to sync their data to the rest of the files stores will use this)* > Note: fileObj must be mounted on a `FS.Collection`, it makes no sense to store otherwise > ```FS.TempStore.createWriteStream = function(fileObj, options) { ...``` [tempStore.js:217](tempStore.js#L217) - ### *fsTempstore*.createReadStream(fileObj)  Server ### *This method __createReadStream__ is defined in `FS.TempStore`* __Arguments__ * __fileObj__ *{[FS.File](#FS.File)}* The file to read __Returns__ *{Stream}* Returns readable stream > ```FS.TempStore.createReadStream = function(fileObj) { ...``` [tempStore.js:313](tempStore.js#L313)