dsfs

package
v0.9.13 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Oct 12, 2020 License: GPL-3.0 Imports: 26 Imported by: 0

Documentation

Overview

Package dsfs glues datsets to cafs (content-addressed-file-system)

Index

Constants

View Source
const (
	// BodyDefault is the default action: compare them to get how much changed
	BodyDefault = BodyAction("default")
	// BodySame means that the bodies are the same, no need to compare
	BodySame = BodyAction("same")
	// BodyTooBig means the body is too big to compare, and should be assumed to have changed
	BodyTooBig = BodyAction("too_big")
)

Variables

View Source
var (
	// BodySizeSmallEnoughToDiff sets how small a body must be to generate a message from it
	BodySizeSmallEnoughToDiff = 20000000 // 20M or less is small
	// OpenFileTimeoutDuration determines the maximium amount of time to wait for
	// a Filestore to open a file. Some filestores (like IPFS) fallback to a
	// network request when it can't find a file locally. Setting a short timeout
	// prevents waiting for a slow network response, at the expense of leaving
	// files unresolved.
	// TODO (b5) - allow -1 duration as a sentinel value for no timeout
	OpenFileTimeoutDuration = time.Millisecond * 700
)
View Source
var BaseTabularSchema = map[string]interface{}{
	"type": "array",
	"items": map[string]interface{}{
		"type":  "array",
		"items": []interface{}{},
	},
}

BaseTabularSchema is the base schema for tabular data NOTE: Do not use if possible, prefer github.com/qri-io/dataset/tabular TODO(dustmop): Possibly move this to tabular package

View Source
var ErrNoReadme = fmt.Errorf("this dataset has no readme component")

ErrNoReadme is the error for asking a dataset without a readme component for readme info

View Source
var ErrNoTransform = fmt.Errorf("this dataset has no transform component")

ErrNoTransform is the error for asking a dataset without a tranform component for viz info

View Source
var ErrNoViz = fmt.Errorf("this dataset has no viz component")

ErrNoViz is the error for asking a dataset without a viz component for viz info

View Source
var Timestamp = func() time.Time {
	return time.Now().UTC()
}

Timestamp is an function for getting commit timestamps timestamps MUST be stored in UTC time zone

Functions

func CreateDataset

func CreateDataset(ctx context.Context, source, destination cafs.Filestore, ds, dsPrev *dataset.Dataset, pk crypto.PrivKey, sw SaveSwitches) (string, error)

CreateDataset places a dataset into the store. Store is where we're going to store the data Dataset to be saved Prev is the previous version or nil if there isn't one Pk is the private key for cryptographically signing Sw is switches that control how the save happens Returns the immutable path if no error

func DerefDataset

func DerefDataset(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDataset attempts to fully dereference a dataset

func DerefDatasetCommit

func DerefDatasetCommit(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDatasetCommit derferences a dataset's Commit element if required should be a no-op if ds.Structure is nil or isn't a reference

func DerefDatasetMeta

func DerefDatasetMeta(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDatasetMeta derferences a dataset's transform element if required should be a no-op if ds.Structure is nil or isn't a reference

func DerefDatasetReadme

func DerefDatasetReadme(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDatasetReadme dereferences a dataset's Readme element if required should be a no-op if ds.Readme is nil or isn't a reference

func DerefDatasetStructure

func DerefDatasetStructure(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDatasetStructure derferences a dataset's structure element if required should be a no-op if ds.Structure is nil or isn't a reference

func DerefDatasetTransform

func DerefDatasetTransform(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDatasetTransform derferences a dataset's transform element if required should be a no-op if ds.Structure is nil or isn't a reference

func DerefDatasetViz

func DerefDatasetViz(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) error

DerefDatasetViz dereferences a dataset's Viz element if required should be a no-op if ds.Viz is nil or isn't a reference

func GetHashBase

func GetHashBase(in, network string) string

GetHashBase strips paths to return just the hash

func JSONFile

func JSONFile(name string, m json.Marshaler) (qfs.File, error)

JSONFile is a convenenience method for creating a file from a json.Marshaller

func LoadBody

func LoadBody(ctx context.Context, store cafs.Filestore, ds *dataset.Dataset) (qfs.File, error)

LoadBody loads the data this dataset points to from the store

func LoadDataset

func LoadDataset(ctx context.Context, store cafs.Filestore, path string) (*dataset.Dataset, error)

LoadDataset reads a dataset from a cafs and dereferences structure, transform, and commitMsg if they exist, returning a fully-hydrated dataset

func LoadDatasetRefs

func LoadDatasetRefs(ctx context.Context, store cafs.Filestore, path string) (*dataset.Dataset, error)

LoadDatasetRefs reads a dataset from a content addressed filesystem without dereferencing it's components

func LoadReadmeScript

func LoadReadmeScript(ctx context.Context, store cafs.Filestore, dspath string) (qfs.File, error)

LoadReadmeScript loads script data from a dataset path if the given dataset has a readme script is specified the returned qfs.File will be the value of dataset.Readme.ScriptPath

func LoadTransformScript

func LoadTransformScript(ctx context.Context, store cafs.Filestore, dspath string) (qfs.File, error)

LoadTransformScript loads transform script data from a dataset path if the given dataset has a transform script specified the returned qfs.File will be the value of dataset.Transform.ScriptPath TODO - this is broken, assumes file is JSON. fix & test or depricate

func PackageFilepath

func PackageFilepath(store cafs.Filestore, path string, pf PackageFile) string

PackageFilepath returns the path to a package file for a given base path It relies relies on package storage conventions and cafs.Filestore path prefixes If you supply a path that does not match the filestore's naming conventions will return an invalid path

func WriteDataset

func WriteDataset(ctx context.Context, destination cafs.Filestore, ds *dataset.Dataset, pin bool) (string, error)

WriteDataset writes a dataset to a cafs, replacing subcomponents of a dataset with path references during the write process. Directory structure is according to PackageFile naming conventions. This method is currently exported, but 99% of use cases should use CreateDataset instead of this lower-level function

Types

type BodyAction added in v0.9.7

type BodyAction string

BodyAction represents the action that should be taken to understand how the body changed

type PackageFile

type PackageFile int

PackageFile specifies the different types of files that are stored in a package

const (
	// PackageFileUnknown is the default package file, which
	// should be erroneous, as there is no sensible default
	// for PackageFile
	PackageFileUnknown PackageFile = iota
	// PackageFileDataset is the main dataset.json file
	// that contains all dataset metadata, and is the only
	// required file to constitute a dataset
	PackageFileDataset
	// PackageFileStructure isolates this dataset's structure
	// in it's own file
	PackageFileStructure
	// PackageFileAbstract is the abstract verion of
	// structure
	PackageFileAbstract
	// PackageFileResources lists the resource datasets
	// that went into creating a dataset
	// TODO - I think this can be removed now that Transform exists
	PackageFileResources
	// PackageFileCommit isolates the user-entered
	// documentation of the changes to this dataset's history
	PackageFileCommit
	// PackageFileTransform isloates the concrete transform that
	// generated this dataset
	PackageFileTransform
	// PackageFileAbstractTransform is the abstract version of
	// the operation performed to create this dataset
	PackageFileAbstractTransform
	// PackageFileMeta encapsulates human-readable metadata
	PackageFileMeta
	// PackageFileViz isolates the data related to representing a dataset as a visualization
	PackageFileViz
	// PackageFileRenderedViz is the rendered visualization of the dataset
	PackageFileRenderedViz
	// PackageFileReadme connects readme data to the dataset package
	PackageFileReadme
	// PackageFileReadmeScript is the raw readme of the dataset
	PackageFileReadmeScript
	// PackageFileRenderedReadme is the rendered readme of the dataset
	PackageFileRenderedReadme
)

func (PackageFile) Filename

func (p PackageFile) Filename() string

Filename gives the canonical filename for a PackageFile

func (PackageFile) String

func (p PackageFile) String() string

String implements the io.Stringer interface for PackageFile

type SaveSwitches added in v0.9.8

type SaveSwitches struct {
	// Replace is whether the save is a full replacement or a set of patches to previous
	Replace bool
	// Pin is whether the dataset should be pinned
	Pin bool
	// ConvertFormatToPrev is whether the body should be converted to match the previous format
	ConvertFormatToPrev bool
	// ForceIfNoChanges is whether the save should be forced even if no changes are detected
	ForceIfNoChanges bool
	// ShouldRender is deprecated, controls whether viz should be rendered
	ShouldRender bool
	// NewName is whether a new dataset should be created, guaranteeing there's no previous version
	NewName bool
	// FileHint is a hint for what file is used for creating this dataset
	FileHint string
	// Drop is a string of components to remove before saving
	Drop string
}

SaveSwitches represents options for saving a dataset

Directories

Path Synopsis
Package dstest defines an interface for reading test cases from static files leveraging directories of test dataset input files & expected output files
Package dstest defines an interface for reading test cases from static files leveraging directories of test dataset input files & expected output files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL