Glossary

This glossary fixes the vocabulary of the Tabalyst documentation. English is the source language; the French terms below are used when Tabalyst is translated, for example on the website.

Never translated

Keep these exactly as written in every language:

• the product names Tabalyst, Tabalyst Report, Tabalyst CSV Report, Tabalyst Scan, Tabalyst Inspect, Tabalyst Dedup and Tabalyst Generate;
• commands and options, such as tabalyst report, -o, --output-dir or --config;
• JSON keys and values, such as format_version, with_issues or mixed;
• file names and paths, such as report.json or executions.json;
• Python modules, functions and parameters, such as tabalyst.analyze().

Terms

English	French	Meaning
toolkit	boîte à outils	What Tabalyst is: a set of tools for unfamiliar data
report	rapport	The interactive HTML file produced by tabalyst report
profile, JSON profile	profil, profil JSON	The report.json file describing a dataset
column profile	profil de colonne	The part of the profile describing one column
dataset	jeu de données	The tabular data being analyzed
source file	fichier source	The CSV file given to Tabalyst
dataset overview	vue d’ensemble du jeu de données	The report section summarizing the whole dataset
finding	constat	What a rule found in a column: a count of one unit, out of a population, with its accuracy. An Error, an Issue or an Info
Error	erreur	A finding that a rule proves invalid, such as a date that does not exist in the calendar
Issue	point d’attention	A finding that is an ambiguity or a deviation worth a look, never proof of an error. The report has no Warning class
signal	signal	A tile of the dataset overview that sums one kind of finding over the columns that have it
classification	classification	The family and the type of a column (for example Identifier / ID, Text / Short), decided from the evidence of the scan and refined by the second pass
Unique ID	identifiant unique	An identifier column with no missing value and an exact count proving that every value is distinct across all the rows (raw uniqueness)
identifier uniqueness	unicité de l’identifiant	The report’s reading of an identifier column that separates the repetitions of the ID that exact copies of whole rows explain from those that remain between distinct records (identifier_uniqueness in the JSON profile)
raw uniqueness	unicité brute	Every value of the ID is distinct across all the source records, copies of rows included; the Unique ID of the Scan
unique after row deduplication	unique après déduplication des lignes	Every value of the ID is distinct among the distinct rows: the only repetitions come from exact copies of whole rows. It is not a Unique ID, which is the raw reading
unique among present IDs	unique parmi les ID présents	The reading of an identifier that lacks a value in some records: after the exact copies of rows are set aside, no ID repeats between the distinct rows that have one. It is never a Unique ID, because the records without an ID remain
near-unique	quasi-unique	A non-binding reading of an identifier whose remaining repetitions between distinct rows are few for their number (at most the smaller of 1% of them, their square root and 1,000). It is never a guarantee and never makes the column a Unique ID
explained repetition	répétition expliquée	A repetition of an ID that an exact extra copy of a whole row accounts for; the unexplained ones are residual repetitions, between distinct records
data sample	échantillon de données	The rows of the row sample shown in the report: the start, sampled rows and the end
raw values	valeurs brutes	Source values before normalization
missing value	valeur manquante	An empty value or a configured missing marker
quality issue	anomalie de qualité	An observation that is a confirmed issue or a candidate anomaly; a remark is not one
with quality issues	avec anomalies de qualité	A column with a confirmed issue or a candidate anomaly, such as an inferred type mixed or ambiguous dates
observation	observation	A finding of a scan about the data, with a status: informational, candidate anomaly or confirmed issue
limit	limite	What a scan could not establish, such as a count stopped by a scan limit; never a defect of the data
value normalization	normalisation des valeurs	Cleaning applied before analysis, such as whitespace trimming
type inference	inférence de type	How Tabalyst decides the type of a column
inferred type	type inféré	The type deduced from the values of a column
physical type	type physique	The storage type of the values: number, date, text
semantic type	type sémantique	The meaning of a column, such as an email or an identifier
date analysis	analyse des dates	The report section on date columns
numeric analysis	analyse numérique	The report section on numeric columns
string analysis	analyse des textes	The report section on text columns
analysis settings	paramètres d’analyse	The effective configuration used for a report
configuration file	fichier de configuration	The JSON file passed with --config
delimiter	séparateur	The character separating CSV fields
encoding	encodage	The character encoding of the source file
execution history	historique d’exécution	The cumulative executions.json file
format revision	révision du format	The format_revision number of the profile format
batch	lot	Several source files processed by one command
output directory	dossier de sortie	The folder given with -d / --output-dir
alpha	alpha	Early stage where interfaces and formats may change
beta	bêta	Pre-1.0 stage where interfaces and formats may still change
scan	analyse	The complete description of a source written by tabalyst scan
scan document	document d’analyse	The project’s scan.json or a standalone .scan.json export written by tabalyst scan
stale scan	analyse périmée	A scan document whose source, settings or pattern discovery changed since it was written
stratum, strata	strate, strates	A group of rows that share the value of the stratification field, or the combination of values of several fields
stratification advice	aide à la stratification	What tabalyst scan --plan --stratify adds to a plan: which columns make good strata and how risky a combination of columns is
smart sample	échantillon smart	A sample that aims to cover the diversity and the rare cases of a file; its proportions do not estimate the population
target	cible	A thing a smart sample aims to cover: a value, a pattern, an extreme, a period, a problem the scan found. It has a weight and is mandatory or optional
smart sample plan	plan d’échantillon smart	What tabalyst scan --plan --smart adds to a plan: the targets, their weights, the sizes, the reserve and the readings of the source
coverage (of a smart sample)	couverture	What a smart sample holds of the targets of its plan, by family and weighted by their weights. It is the coverage of the catalog of targets, not of every phenomenon of the data
reserve	réserve	The ordinary rows of a smart sample, chosen by a seeded hash of their position, so that it does not hold only rare cases. It gives way to mandatory targets
distinct sample	échantillon de valeurs distinctes	The distinct original values of one field kept in values.samples, all of them or the bottom-k of smallest priority, each with its exact count over the whole field. It is not a ranking
row sample	échantillon de lignes	Whole rows of a dataset kept in records.sample (the start, the end and a reproducible choice of the interior), so that two fields can be related on the same rows. It is not a uniform sample of the file
estimated figure	valeur estimée	A figure with accuracy: "estimated", such as the KMV estimate of a distinct count, with its expected error. It is never quoted as a count
limited result	résultat limité	A measure stopped by a configured limit, written with its reason and a proven lower_bound, never with an estimate in place of the value
fingerprint-based identity	identité par empreinte	identity_accuracy: "fingerprint_based": two long texts are the same entry when their 128-bit digests are equal, not when the full texts were compared
analysis plan	plan d’analyses	What tabalyst scan --plan shows: the analyses a scan could receive, their cost and the decision on each
forced plan	plan forcé	A plan that takes every applicable analysis within the hard memory budget, whatever the limits of an automatic choice (tabalyst scan --plan --deep)
confirmation	confirmation	The question a costly run asks before it starts; a plan says whether it would, and why
second pass	second passage	The reading of a source after its scan, to add analyses such as histograms, patterns counted on every value and dates per period (tabalyst scan --deep, and by itself in tabalyst report); its result is the second pass document
validated pattern	pattern validé	An inferred pattern whose occurrences were counted on every value of the file by the second pass, not only on a sample of the distinct values
format	format	The exact form of one value, a symbol per character (AA-9999)
pattern	pattern	The values of one structure and one set of lengths, with their constants ('AK-'9999); a discovered pattern never has a range
family	famille	The patterns of one structure that differ only by lengths or constants; shown as the list of its patterns with their counts
character map	carte des caractères	The most frequent characters of a field, their scripts and Unicode categories, the case of the ends of the values and the emoji; counted on the scan’s samples, and on every value by the second pass
facet	facette	A dimension along which a family is specialized: composition, lengths, or the content of a token (head, tail, whole)
node	nœud	A pattern plus a constraint on a facet, with its counts in distinct values and occurrences
deep.mode	mode du second passage	The setting that tells tabalyst report whether to plan a second pass by itself (auto), never (off) or always, as --deep does (force)
second pass document	document de second passage	The deep.json or <stem>.deep.json written by tabalyst scan --deep, beside the scan, which it never changes
pair of columns	paire de colonnes	Two numeric columns of one dataset analyzed together. The planner scores the pairs on the row sample, lists the best and the second pass computes their correlation over every row
common rows	lignes communes	The rows where both columns of a pair have a number; the means and the coefficient of a pair are those of these rows only (pairwise_complete)
undefined coefficient	coefficient non défini	A correlation that cannot be computed, because too few common rows or a column that does not vary; the result says why and never writes 0
category	catégorie	A column of a few repeated values, such as a region, whose value is the row of a table; its exact frequencies come from the scan
correlation ratio	rapport de corrélation	The share of the variance of a number that the categories of another column explain, from 0 to 1 (eta_squared). The second pass also gives it corrected for the number of categories (epsilon_squared), and the planner rates a pair by its square root
skewness	asymétrie	Whether the long tail of a numeric column is on the right (positive) or on the left (negative); the third standardized moment. The second pass gives it with the excess kurtosis, which grows with the weight of the tails
outlier of Tukey	valeur aberrante de Tukey	A number beyond a fence at 1.5 (mild) or 3 (far) interquartile ranges below the first quartile or above the third; it is a candidate anomaly to look at, not an error
feature	caractéristique dérivée	What a table reads of a column to make its classes: the pattern that reads the value, its length, its prefix or its value (another category)
table of a category against a feature	table croisée	The exact count of the rows of each category in each class of a feature of another column, computed by the second pass over every row (category_crosstab)
class	classe	A column of a table: a pattern, a length or a prefix; the last one is always other, what the others do not take
purity	pureté	The share of the classified rows of a category that sit in its dominant class; the purity of a table is that of all its categories together, and it differs from category to class and from class to category
exception	exception	A row outside the dominant class of its category, counted exactly and shown with its first rows
rank correlation	corrélation de rang	The Spearman coefficient of a pair: the Pearson coefficient of the ranks, which sees a relation that always goes up without being a line. The second pass estimates it from a grid of bins and gives certain bounds
provided rule	règle fournie	A rule you wrote for a table, such as “region east takes plan basic”, checked as it is against the counts with its counterexamples; it is never proposed by itself
candidate rule	règle candidate	A relation learned from the data, such as “region QC reads pattern XX9999”, that nobody validated; it is not a rule of the data
Cramér’s V	V de Cramér	The strength of the association of two columns in a table, from 0 to 1, computed from the counts; the corrected form does not grow with the number of classes and is what Tabalyst ranks by. It is not defined, never 0, when a single category or class has rows
exact distinct count	nombre exact de valeurs distinctes	The number of different original values of a field, counted exactly by the second pass when the scan only estimated it; redundant is the occurrences beyond the first of each value
insight card	carte d’analyse	What a Report prints for one analysis of the second pass and one column: figures, distributions and tables, each with its own accuracy label (exact, estimated, sample, at least). It is derived from the document when the page is rendered, never stored
slice	tranche	One tenth of the records of a dataset, in file order, with the same number of records (one more or fewer): the unit in which the second pass follows a column along the file
drift along the file	dérive selon la position	A change of the missing values or of the values of a column from one slice of the file to the next, counted exactly; it shows that the column was filled or exported differently, not why
set of keys	ensemble de clés	The keys an object of a JSON file has; objects with the same keys are one set, counted exactly by the second pass
fingerprint	empreinte	A 128-bit digest that tells two long values apart; two values are taken for one only if their digests collide, which is about as unlikely as 2^-128
mask	masque	A value with each letter shown as a or A and each digit as 9; the form of a sensitive value that the document may hold
floor (forms)	seuil	The number of values (deep.sensitive.min_count) under which a mask or a class is summed up, so no figure is given about so few values
missing together	manquants simultanés	Columns that lack a value in the same records, counted exactly by the second pass; the share where both do among the records where either does ranks the pairs
histogram	histogramme	The count of the values of a numeric field per integer or per bin, exact, with a median and quantiles that are exact or estimated
already satisfied	déjà satisfaite	An analysis whose answer the scan already holds, so the plan does not select it
duplicate record	enregistrement en double	A record equal to an earlier record of its dataset
row identity	identité de ligne	In Dedup, the list of the decoded cells of a CSV record, compared without any normalization: case, spaces and leading zeros count
distinct row	ligne distincte	One row identity; Dedup counts it once, and writes it once to unique.csv
duplicate group	groupe de doublons	All the occurrences of one row that occurs at least twice; records.duplicate_groups counts the groups
repeated row	ligne répétée	The row of a duplicate group, written once to duplicates.csv
extra copy	occurrence supplémentaire	An occurrence after the first of its row, counted by records.extra_copies; a row that occurs three times has two
first occurrence	première occurrence	The first record of a row in the order of the source; the one Dedup keeps
singleton	ligne unique	A row that occurs exactly once
maximum frequency	fréquence maximale	records.max_occurrences, the most occurrences of one row; three or more makes Dedup write duplicates_all.csv
group occurrences	occurrences des groupes	Every occurrence of the rows of the duplicate groups, first ones included; the rows of duplicates_all.csv
managed output	sortie gérée	A file under an output name of Dedup that a Dedup report proves it published
foreign output	sortie étrangère	A file under an output name of Dedup that no report proves it published; Dedup replaces it only with --force
recovery report	rapport de reprise	The report with the status in_progress that Dedup writes before it changes the files of an earlier result
verification level	niveau de vérification	metadata (size and modification time) or digest (SHA-256): how far Dedup checked that a skipped source is unchanged
storage mode	mode de stockage	memory or disk: where Dedup kept the rows it had seen
spill	repli sur disque	The automatic switch of Dedup from memory to disk when the memory budget is reached
record	enregistrement	One CSV row, one element of a JSON collection, one object line of a JSONL file or one data row of an Excel table
collection	collection	A JSON array whose elements are analyzed as records; a JSONL file is one collection, $[]
field	champ	A CSV column or a path inside JSON records, such as orders[].amount
detector	détecteur	A rule that recognizes a kind of value, such as email addresses or dates
interpretation	interprétation	A meaning proposed for a field by the detectors that match its values
sensitive value	valeur sensible	A value identifying a person, such as an email address; shown by default, maskable
mask	masque	A value with letters replaced by A or a and digits by 9
measure envelope	enveloppe de mesure	The status wrapper saying whether a measure is complete or limited
diagnostic	diagnostic	A technical event of a scan, such as an excluded record or a reached limit
configuration layer	couche de configuration	One level of settings: defaults, each configuration file, then options
JSONL	JSONL	A file with one JSON object per line, ending in .jsonl or .ndjson
workbook	classeur	An Excel file, .xlsx or .xlsm, with one or more sheets
sheet	feuille	One tab of a workbook; a candidate table of its own unless it holds named tables
named table	tableau nommé	An Excel table (list object) defined in a sheet, with its own header and range
header row	ligne d’en-tête	The row of a sheet whose cells name the columns, detected or set by excel.header_row
inspect, Tabalyst Inspect	inspection, Tabalyst Inspect	The step that reads a JSON, JSONL or Excel source, finds its collection or table and writes the Inspect file; the command is tabalyst inspect
Inspect file	fichier Inspect	The <source>-inspect.json file beside a source, with the detection and the config you may edit
detection	détection	The section of an Inspect file that describes what Inspect found
candidate collection	collection candidate	An array, or for a workbook a sheet or named table, that Inspect could analyze as the dataset of a source
eligible candidate	candidat éligible	A candidate collection that has at least one element, all of them objects
selection	sélection	The collection Inspect proposes, or none
ambiguous selection, unresolved selection	sélection ambiguë, sélection non résolue	Several collections are equally plausible, or none is usable, so nothing is selected and tabalyst scan and tabalyst report stop
flatten	aplatissement	Reading nested objects as fields named by their path, such as address.city
complex value	valeur complexe	An object or array kept whole instead of being developed into fields
error policy	politique d’erreurs	strict or tolerant: whether a malformed record stops the scan or is excluded
partial scan	analyse partielle	A scan with status partial, which excluded some records under the tolerant policy