Module `tf.cheatsheet`

`A.` Advanced API

Initialization, configuration, meta data, and linking

A = use('org/repo'): start up and load a corpus from a repository and deliver its API.; See tf.about.usefunc
A.hoist(globals()): Make the API handles F, E, T, L etc available in the global scope.; App.load()
A.load(features): Load an extra bunch of features.; App.load()
A.featureTypes(show=True): show for which types each feature is defined; App.featureTypes()
A.showContext(...): show app settings; showContext()
A.header(allMeta=False): show colophon; header()
A.showProvenance(...): show provenance of code and data; showProvenance()
A.webLink(n, ...): hyperlink to node n on the web; webLink()
A.flexLink("pages") A.flexLink("tut"): hyperlink to app tutorial and documentation; flexLink()
A.isLoaded(features=None): Show information about loaded features; Api.isLoaded()
A.footprint(): Show memory footprint per feature; Api.footprint()

Displaying

A.specialCharacters(): show all hard-to-type characters in the corpus in a widget; specialCharacters()
A.showFormats(): show all text formats and their definitions; showFormats()
A.dm(markdownString): display markdown string in notebook; dm()
A.dh(htmlString): display HTML string in notebook; dh()
A.method(option1=value1, option2=value2, ...): Many of the following methods accept these options as keyword arguments:; tf.advanced.options
A.displayShow(...): show display options; displayShow()
A.displayReset(...): reset display options; displayReset()
A.displaySetup(...): set up display options; displaySetup()
A.table(results, ...): plain rendering of tuple of tuples of node; table()
A.plainTuple(tup, ...): plain rendering of tuple of node; plainTuple()
A.plain(node, ...): plain rendering of node; plain()
A.show(results, ...): pretty rendering of tuple of tuples of node; show()
A.prettyTuple(tup, ...): pretty rendering of tuple of node; prettyTuple()
A.pretty(node, ...): pretty rendering of node; pretty()
A.unravel(node, ...): convert a graph to a tree; unravel()
A.getCss(): get the complete CSS style sheet for this app; getCss()

Search (high level)

A.search(...): search, collect and deliver results, report number of results; search()

Sections and Structure

A.nodeFromSectionStr(...): lookup node for section heading; nodeFromSectionStr()
A.sectionStrFromNode(...): lookup section heading for node; sectionStrFromNode()
A.structureStrFromNode(...): lookup structure heading for node; structureStrFromNode()

Volumes and collections

Export to Excel

A.export(results, ...): export formatted data; export()

Logging

A.dm(markdownString): display markdown string in notebook; dm()
A.dh(htmlString): display HTML string in notebook; dh()
A.version: version number of data of the corpus.; Fabric.version

The following methods work also for TF. instead of A.:

A.banner: banner of the TF program.; Fabric.banner
A.isSilent(): report the verbosity of TF; Timestamp.isSilent()
A.silentOn(deep=False): make TF (deeply) silent from now on.; Timestamp.silentOn()
A.silentOff(): make TF talkative from now on.; Timestamp.silentOff()
A.setSilent(silent): set the verbosity of TF.; Timestamp.setSilent()
A.indent(level=None, reset=False): Sets up indentation and timing of following messages; Timestamp.indent()
A.info(msg, tm=True, nl=True, ...): informational message; Timestamp.info()
A.warning(msg, tm=True, nl=True, ...): warning message; Timestamp.warning()
A.error(msg, tm=True, nl=True, ...): error message; Timestamp.error()

`N. F. E. L. T. S. C.` Core API

`N.` Nodes

Read about the canonical ordering here: tf.core.nodes.

N.walk(): generator of all nodes in canonical ordering; Nodes.walk()
N.sortNodes(nodes): sorts nodes in the canonical ordering; Nodes.sortNodes()
N.otypeRank[nodeType]: ranking position of nodeType; Nodes.otypeRank
N.sortKey(node): defines the canonical ordering on nodes; Nodes.sortKey
N.sortKeyTuple(tup): extends the canonical ordering on nodes to tuples of nodes; Nodes.sortKeyTuple
N.sortKeyChunk(node): defines the canonical ordering on node chunks; Nodes.sortKeyChunk

`F.` Node features

Fall(): all loaded feature names (node features only); Api.Fall()
F.fff.v(node): get value of node feature fff; NodeFeature.v()
F.fff.s(value): get nodes where feature fff has value; NodeFeature.s()
F.fff.freqList(...): frequency list of values of fff; NodeFeature.freqList()
F.fff.items(...): generator of all entries of fff as mapping from nodes to values; NodeFeature.items()
F.fff.meta: meta data of feature fff; NodeFeature.meta
Fs('fff'): identical to F.ffff, usable if name of feature is variable; Api.Fs()

Special node feature `otype`

Maps nodes to their types.

F.otype.v(node): get type of node; OtypeFeature.v()
F.otype.s(nodeType): get all nodes of type nodeType; OtypeFeature.s()
F.otype.sInterval(nodeType): gives start and ending nodes of nodeType; OtypeFeature.sInterval()
F.otype.items(...): generator of all (node, type) pairs.; OtypeFeature.items()
F.otype.meta: meta data of feature otype; OtypeFeature.meta
F.otype.maxSlot: the last slot node; OtypeFeature.maxSlot
F.otype.maxNode: the last node; OtypeFeature.maxNode
F.otype.slotType: the slot type; OtypeFeature.slotType
F.otype.all: sorted list of all node types; OtypeFeature.all

`E.` Edge features

Eall(): all loaded feature names (edge features only); Api.Eall()
E.fff.f(node): get value of feature fff for edges from node; EdgeFeature.f()
E.fff.t(node): get value of feature fff for edges to node; EdgeFeature.t()
E.fff.freqList(...): frequency list of values of fff; EdgeFeature.freqList()
E.fff.items(...): generator of all entries of fff as mapping from edges to values; EdgeFeature.items()
E.fff.b(node): get value of feature fff for edges from and to node; EdgeFeature.b()
E.fff.meta: all meta data of feature fff; EdgeFeature.meta
Es('fff'): identical to E.fff, usable if name of feature is variable; Api.Es()

Special edge feature `oslots`

Maps nodes to the set of slots they occupy.

E.oslots.items(...): generator of all entries of oslots as mapping from nodes to sets of slots; OslotsFeature.items()
E.oslots.s(node): set of slots linked to node; OslotsFeature.s()
E.oslots.meta: all meta data of feature oslots; OslotsFeature.meta

`L.` Locality

L.i(node, otype=...): go to intersecting nodes; Locality.i()
L.u(node, otype=...): go one level up; Locality.u()
L.d(node, otype=...): go one level down; Locality.d()
L.p(node, otype=...): go to adjacent previous nodes; Locality.p()
L.n(node, otype=...): go to adjacent next nodes; Locality.n()

`T.` Text

T.text(node, fmt=..., ...): give formatted text associated with node; Text.text()

Sections

Rigid 1 or 2 or 3 sectioning system

T.sectionTuple(node): give tuple of section nodes that contain node; Text.sectionTuple()
T.sectionFromNode(node): give section heading of node; Text.sectionFromNode()
T.nodeFromSection(section): give node for section heading; Text.nodeFromSection()

Structure

Flexible multilevel sectioning system

T.headingFromNode(node): give structure heading of node; Text.headingFromNode()
T.nodeFromHeading(heading): give node for structure heading; Text.nodeFromHeading()
T.structureInfo(): give summary of dataset structure; Text.structureInfo()
T.structure(node): give structure of node and all in it.; Text.structure()
T.structurePretty(node): pretty print structure of node and all in it.; Text.structurePretty()
T.top(): give all top-level structural nodes in the dataset; Text.top()
T.up(node): gives parent of structural node; Text.up()
T.down(node): gives children of structural node; Text.down()

`S.` Search (low level)

searchRough

Preparation

S.search(query, limit=None): Query the TF dataset with a template; Search.search()
S.study(query, ...): Study the query in order to set up a plan; Search.study()
S.showPlan(details=False): Show the search plan resulting from the last study.; Search.showPlan()
S.relationsLegend(): Catalog of all relational devices in search templates; Search.relationsLegend()

Fetching results

S.count(progress=None, limit=None): Count the results, up to a limit; Search.count()
S.fetch(limit=None, ...): Fetches the results, up to a limit; Search.fetch()
S.glean(tup): Renders a single result into something human readable.; Search.glean()

Implementation

S.tweakPerformance(...): Set certain parameters that influence the performance of search.; Search.tweakPerformance()

`C.` Computed data components.

Access to pre-computed data: Computeds.

All components have just one useful attribute: .data.

Call(): all pre-computed data component names; Api.Call()
Cs('ccc'): identical to C.ccc, usable if name of component is variable; Api.Cs()
C.levels.data: various statistics on node types; levels()
C.order.data: the canonical order of the nodes (tf.core.nodes); order()
C.rank.data: the rank of the nodes in the canonical order (tf.core.nodes); rank()
C.levUp.data: feeds the Locality.u() function; levUp()
C.levDown.data: feeds the Locality.d() function; levDown()
C.boundary.data: feeds the Locality.p() and Locality.n() functions; boundary()
C.characters.data: frequency list of characters in a corpus, separately for all the text formats; characters()
C.sections.data["sec1"] C.sections.data["sec2"]: feeds the section part of tf.core.text; sections()
C.sections.data["seqFromNode"] C.sections.data["nodeFromSeq"]: maps tuples of heading nodes to their corresponding tuples of sequence numbers and vice versa. Only if there are 3 section levels.; sections()
C.structure.data: feeds the structure part of tf.core.text; structure()

`TF.` Dataset

Loading

TF = Fabric(locations=dirs, modules=subdirs, volume=None, collection=None, silent="auto"): Initialize API on work or single volume or collection of a work from explicit directories. Use use() instead wherever you can. See also tf.about.volumes.; Fabric
TF.isLoaded(features=None): Show information about loaded features; Api.isLoaded()
TF.explore(show=True): Get features by category, loaded or unloaded; FabricCore.explore()
TF.loadAll(silent="auto"): Load all loadable features.; FabricCore.loadAll()
TF.load(features, add=False): Load a bunch of features from scratch or additionally.; FabricCore.load()
TF.ensureLoaded(features): Make sure that features are loaded.; Api.ensureLoaded()
TF.makeAvailableIn(globals()): Make the members of the core API available in the global scope; Api.makeAvailableIn()
TF.ignored: Which features have been overridden.; Api.ignored
TF.footprint(): Show memory footprint per feature; Api.footprint()

Volumes

Saving and Publishing

TF.save(nodeFeatures={}, edgeFeatures={}, metaData={},,...): Save a bunch of newly generated features to disk.; FabricCore.save()
A.publishRelease(increase, message=None, description=None,,...): Commit the dataset repo, tag it, release it, and attach the complete zipped data to it.; publishRelease()

Export to ZIP

cd ~/backend/org/repo tf-zipall: store the complete corpus data in a file complete.zip; zipAll()
A.zipAll(): store the complete corpus data in a file complete.zip; zipAll()
from tf.app import collect collect(backend, org, repo): same as A.zipAll() above, assuming the data is in a Github clone; collect()

House keeping

TF.version: version number of TF.; Fabric.version
TF.clearCache(): clears the cache of compiled TF data; FabricCore.clearCache()

from tf.clean import clean

clean(): clears the cache of compiled TF data; tf.clean

Volume support

TF datasets per volume or collection of a work. See also tf.about.volumes.

from tf.volumes import getVolumes getVolumes(volumeDir): List volumes in a directory.; getVolumes()
from tf.volumes import extract extract(work, volumes, ...): Extracts volumes from a work; tf.volumes.extract
from tf.volumes import collect collect(volumes, work, ...): Collects several volumes into a new collection; tf.volumes.collect

Dataset Operations

from tf.dataset import modify modify(source, target, ...): Modifies a TF dataset into a new TF dataset; tf.dataset.modify
from tf.dataset import Versions Versions(api, va, vb, slotMap): Extends a slot mapping between versions of a TF dataset to a complete node mapping; tf.dataset.nodemaps

Data Interchange

Custom node sets for search

from tf.lib import readSets
from tf.lib import writeSets

readSets(sourceFile): reads a named sets from file; readSets()
writeSets(sets, destFile): writes a named sets to file; writeSets()

Export to Excel

A.export(results, ...): export formatted data; export()

Interchange with external annotation tools

from convert.recorder import Recorder

Recorder(): generate annotatable plain text and import annotations; tf.convert.recorder

`pandas` export

A.exportPandas(): export dataset as pandas data frame; tf.convert.pandas

MQL interchange

TF.exportMQL(mqlDb, exportDir=None) A.exportMQL(mqlDb, exportDir=None): export loaded dataset to MQL; exportMQL()
from tf.convert.mql import importMQL TF = importMQL(mqlFile, saveDir): convert MQL file to TF dataset; importMQL()

Walker conversion

from tf.convert.walker import CV

cv = CV(TF): convert structured data to TF dataset; tf.convert.walker

Exploding

from tf.convert.tf import explode

explode(inLocation, outLocation): explode TF feature files to straight data files without optimizations; explode()

TF App development

A.reuse(): reload configuration data; App.reuse()

from tf.advanced.find import loadModule

mmm = loadModule("mmm", *args): load specific module supporting the corpus app; loadModule()
~/mypath/myname/app/config.yaml: settings for a TF App; tf.advanced.settings

Layered search

(these work on the command-line if TF is installed)

tf-make {dataset} {client} ship: generate a static site with a search interface in client-side JavaScript and publish it to GitHub pages. If {client} is left out, generate all clients that are defined for this dataset. Clients are defined in the app-{dataset} repo, under layeredsearch. More commands here.; tf.client.make.build
tf-make {dataset} serve: serve the search interfaces defined for {dataset} locally.

More commands here.

Annotation tools

(these work in the TF browser and in Jupyter Notebooks)

Named Entity Annotation

tf {org}/{repo} --tool=ner: Starts the TF browser for the corpus in org/repo and opens the manual annotation tool.; tf.about.annotateBrowser
NE = A.makeNer(): Sets up the 'manual' annotation API for the corpus in A.; tf.ner.ner; More info and examples in tf.about.annotate .

Command-line tools

(these work on the command-line if TF is installed)

tf {org}/{repo} tf {org}/{repo}: Starts the TF browser for the corpus in org/repo.; tf.browser.start
tf-zipall: Zips the TF dataset located by the current directory, with all its additional data modules, but only the latest version, so that it can be attached to a release on GitHub / GitLab.; zipAll() and tf.zip
tf-zip {org}/{repo}: Zips the TF dataset in org/repo so that it can be attached to a release on GitHub / GitLab.; tf.advanced.zipdata
tf-nbconvert {inDirectory} {outDirectory}: Converts notebooks in inDirectory to HTML and stores them in outDirectory.; tf.tools.nbconvert

A. Advanced API

Initialization, configuration, meta data, and linking

Displaying

Search (high level)

Sections and Structure

Volumes and collections

Export to Excel

Logging

N. F. E. L. T. S. C. Core API

N. Nodes

F. Node features

Special node feature otype

E. Edge features

Special edge feature oslots

L. Locality

T. Text

Sections

Structure

S. Search (low level)

Preparation

Fetching results

Implementation

C. Computed data components.

TF. Dataset

Loading

Volumes

Saving and Publishing

Export to ZIP

House keeping

Volume support

Dataset Operations

Data Interchange

Custom node sets for search

Export to Excel

Interchange with external annotation tools

pandas export

MQL interchange

Walker conversion

Exploding

TF App development

Layered search

Annotation tools

Named Entity Annotation

Command-line tools

`A.` Advanced API

`N. F. E. L. T. S. C.` Core API

`N.` Nodes

`F.` Node features

Special node feature `otype`

`E.` Edge features

Special edge feature `oslots`

`L.` Locality

`T.` Text

`S.` Search (low level)

`C.` Computed data components.

`TF.` Dataset

`pandas` export