Settings#
This page is generated from the code and must not be edited by hand.
A project’s settings.yaml holds only the values a person chose. Every setting not named there takes the default shown here. A dotted name is a nested key: hash.mode is the key mode inside the key hash.
There are 168 settings in 15 groups. The Dashboard shows the same list on its Engine tab, with the same words. Copy puts the setting, at its default, on your clipboard as the lines of settings.yaml that hold it: change the value and paste it into the file. If the file already has the group (for example scan:), put the copied line under that group and leave out the group’s name from the copy. A file that names one group twice is read without an error, and only the last of them counts.
How deep a run reads#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
|
How far a dive goes unless told: never open files, standard, or deep |
Walking the folders#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
a list of words |
Folder and file name patterns the walk skips, with everything beneath them |
|
|
|
whole number, at least 1 |
How many records gather before they are written to disk |
|
|
|
number, more than 0 |
Longest time records wait before being written to disk |
|
|
on |
on or off |
Whether a file that moved is recognized, keeping its readings and decisions |
|
|
off |
fixed; it cannot be changed |
Whether shortcuts to other places are followed; always off, cannot be changed |
|
|
off |
on or off |
Whether hidden files and folders, whose names start with a dot, are walked |
|
|
off |
on or off |
Whether earlier Ingence projects and runs found inside the data are indexed as data |
|
|
on |
on or off |
Whether an unchanged file keeps what an earlier dive read of it |
|
|
not set |
whole number, or left empty |
How many folders deep the walk goes; blank means no limit |
|
|
|
whole number, at least 1 |
Past this many files alike in name, size and time, look-alikes are not paired for moves |
|
|
|
whole number, at least 1 |
Fewest of a folder’s files that must arrive together to count the folder as moved |
|
|
|
number, more than 0, at most 1 |
Share of a folder’s contents that must match one to one to count it moved |
|
|
on |
on or off |
Whether, on one disk, the disk’s own number for each file decides move or copy |
|
|
on |
on or off |
Whether a dive checks a carried move by comparing the file’s hash |
|
|
off |
on or off |
Whether a copy, its original removed, starts afresh instead of taking the original’s readings and answers |
|
|
on |
on or off |
Whether the walk stays on one disk, never crossing into other mounted drives |
|
|
on |
on or off |
Whether a shortcut’s target is checked to see if it still exists |
|
|
|
whole number, from 1 to 256 |
How many folders the walk lists at once |
Comparing files#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
text |
The method used to hash file contents |
|
|
|
whole number, at least 4096 |
How much of a file is read at a time while hashing |
|
|
empty list |
a list of words |
File extensions never hashed |
|
|
|
whole number, at least 0 |
Largest file hashed whole; bigger ones are only sampled, so never proven identical |
|
|
empty list |
a list of words |
When filled, only these file extensions are hashed |
|
|
not set |
whole number, or left empty |
Files larger than this are never hashed; blank means no limit |
|
|
|
|
What a dive hashes: only what comparing files needs, every file, or none |
|
|
|
|
What ‘need’ goes by: what comparing files needs, or the dive’s depth |
|
|
not set |
none: left to the depth; |
Hashing across network shares: left to the depth, always done, or never done |
Reading inside files#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
empty list |
a list of words |
Which file readers may run; empty allows every available one |
|
|
|
|
When files inside installed programs are read: after everything else, never, or in turn |
|
|
empty list |
a list of words |
Dictionary sections whose words are searched for inside file contents; empty means none |
|
|
|
a list of words |
File extensions read only for their timestamp when the ND2 clock-only option is on |
|
|
not set |
number, or left empty |
Total time a dive may spend reading inside files; each dive’s depth sets it |
|
|
off |
on or off |
Whether a dive reads inside files at all; each dive’s depth sets it |
|
|
empty list |
a list of words |
File extensions whose contents are never read |
|
|
on |
on or off |
Whether a reading is also discarded if the file was replaced while being read |
|
|
empty list |
a list of words |
When filled, only these file extensions are read inside |
|
|
|
whole number, at least 0 |
How many levels of archives within archives may be opened |
|
|
|
whole number, at least 0 |
Most files one archive may yield when opened |
|
|
|
number, more than 0 |
How many times its own size an archive’s contents may grow to when opened |
|
|
not set |
whole number, or left empty |
Files larger than this are not read inside; blank means no limit |
|
|
|
whole number, at least 0 |
How much of one file a reading may take in at most; each dive’s depth sets it |
|
|
|
whole number, at least 0 |
Most characters of text kept from one file |
|
|
on |
on or off |
Whether a large file’s header is read first, and its body only if still needed |
|
|
off |
on or off |
Whether ND2 files are read only as far as their header and timestamp |
|
|
on |
on or off |
Whether files found inside archives become items of their own |
|
|
|
|
The order a dive reads files: file types in turn, extensions in turn, or smallest first |
|
|
|
whole number, or left empty |
Memory ceiling for a reading run apart from the engine; blank means none |
|
|
|
number, at least 0 |
Extra seconds of reading time allowed per gigabyte of file |
|
|
|
number, more than 0 |
Seconds one file’s reading may run before it is stopped, plus the per-gigabyte allowance |
|
|
|
whole number, at least 0, or left empty |
Above this size, every reading runs apart so its time limit applies; blank means never |
|
|
|
whole number, at least 1024 |
How much of a text file is read to spot near-duplicates |
|
|
on |
on or off |
Whether readers that ask for it run apart, so one bad file cannot hang the run |
Reading names#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
on |
on or off |
Whether the Dictionary is asked what each part of a name means |
|
|
|
number, more than 1 |
How many times longer than nearby pauses a pause must be to end a burst |
|
|
|
whole number, at least 3 |
How many neighboring pauses the typical pause is judged against |
|
|
on |
on or off |
Whether spellings joined at a capital are read by evidence, rather than always as pieces |
|
|
|
|
For a one-word spelling the Dictionary lacks: treat it as unknown, or read its known pieces |
|
|
|
number, more than 0 |
How often a joined spelling must appear, against its rarest piece alone, to be one word |
|
|
on |
on or off |
Whether each folder’s name patterns, time span and bursts of activity are described |
|
|
|
|
Whether sibling short dates together settle an ambiguous day, month and year order |
|
|
|
|
Whether every shipped domain pack loads (but those refused), or only the ones named |
|
|
empty list |
a list of words |
The shipped domain packs to load when only named ones are chosen |
|
|
|
whole number, at least 0 |
How many files in an instrument’s format (ND2 for microscopy) accept its pack without asking, 0 never. Once a pack is accepted, the packs the project chose (accepted ones and the laboratory’s own) stand above the core and every other shipped pack level with it: a word they spell differently is asked once |
|
|
|
|
How letters and digits written together are read: by what leads, as one word, or split |
|
|
|
whole number, at least 1 |
The fewest letters each part may have when a run of two joined known words is divided |
|
|
|
|
Whether a run of two known words written together, like channelhoechst, is divided |
|
|
|
whole number, at least 1 |
How many pairs like sub-01 a name needs before each reads as a numbered label |
|
|
empty list |
a list of words |
Vocabulary packs the lab installed that a run loads, in order |
|
|
|
a list of numbers |
Older year range for dates in names; kept so old projects load, no longer used |
|
|
|
whole number, at least 2 |
The fewest letters a word needs before it is read as the plural of a known word |
|
|
on |
on or off |
Whether where each part sat within a name is recorded Read by the rules |
|
|
|
|
Whether six digits alone are read as a year-month-day date or as a number |
|
|
|
whole number, at least 1 |
How many known words are offered beside a word that may be a misspelling |
|
|
|
whole number, from 1 to 3 |
How many typing slips apart a word may be from a known word it is offered as |
|
|
|
whole number, at least 3 |
The fewest letters a word needs before a known word is offered as its spelling |
|
|
on |
on or off |
Whether every name is split into words and the result recorded |
|
|
|
a list of years or the word present |
The years a date in a name may fall in, first to last or the present |
The run’s journal#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
on |
on or off |
Whether a run keeps a detailed journal of its settings, timings and decisions |
|
|
on |
on or off |
Whether the journal records one line for each file a dive reads |
|
|
on |
on or off |
Whether the journal records how long each step took for each file |
Archive listings#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
|
Whether links an archive’s listing declares enter the Index, or are held back for scoring |
|
|
|
number, from 0 to 60 |
Seconds to wait between two requests to one archive |
|
|
off |
on or off |
Whether to ask an archive for each file’s date where its listing gives none |
|
|
|
text |
What each request to an archive says it comes from |
Background jobs#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
|
Whether background jobs run as their own program or inside the Dashboard |
|
|
|
number, more than 0, at most 60 |
How often a running job says it is still working |
|
|
|
whole number, from 1 to 100000 |
How many jobs’ logs are kept before older ones are removed |
|
|
|
whole number, from 0 to 20 |
How many times an interrupted job is tried again before it is marked failed |
|
|
|
number, more than 0, at most 3600 |
How long a job may go unheard before it is reported silent |
|
|
on |
on or off |
Whether a job silent too long is stopped and tried again |
|
|
|
number, more than 0, at most 86400 |
How long a job must be silent before it is stopped |
Applying the rules#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
whole number, at least 1 |
How many items are held at once when rules are applied folder by folder |
|
|
|
number, at least 0, less than 1 |
Share of the machine’s memory left free when deciding whether to go folder by folder |
|
|
|
whole number, at least 1 |
Estimated memory one item needs, used to decide whether to go folder by folder |
|
|
|
|
Whether rules are applied folder by folder to save memory: when needed, always, or never |
Completeness and no analysis found#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
number, more than 0, at most 1 |
Reason ‘forgotten’: the share of the rest of its theme’s raw data found analyzed, at least |
|
|
|
whole number, from 1 to 1000000 |
How many raw acquisitions an experiment needs found before it can be listed as having no analysis found |
|
|
|
number, more than 0, at most 1 |
Below this share of its raw acquisitions found followed by analysis, an experiment is listed as having no analysis found |
|
|
|
whole number, from 1 to 1000000 |
How many analysis files found count as an analysis of the experiment’s own |
|
|
|
whole number, from 1 to 1000000 |
How many results tables found count as an analysis of the experiment’s own |
|
|
|
number, at least 0, less than 1 |
Reason ‘a line that stopped’: the share of the rest of its theme’s raw data found analyzed, under |
The design grid of an experiment#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
whole number, from 1 to 1000000 |
The most combinations of an experiment’s grid listed as not found; a larger grid is counted and not listed |
|
|
off |
on or off |
Off by default while it is measured; turn it on to see a grid of the conditions an experiment’s names vary |
|
|
|
whole number, from 2 to 1000000 |
The most different values a part of a name may take and still be drawn as a condition; a part with more is taken for an identifier or a frame number |
|
|
off |
on or off |
Whether a file that no group of repeats holds, such as the only file of its condition, is drawn in the grid when its name is laid out like the others; on, scripts and documents with alike names make grids too |
|
|
|
number, from 0 to 0.5 |
The share of an experiment’s files that may break the pairing of two parts of a name that go together, such as GFP and 488 |
|
|
|
whole number, from 1 to 1000000 |
How many files each value of a part needs before the part may be joined with another that always goes with it |
|
|
on |
on or off |
Whether two parts of a name that always go together, such as GFP and 488, are drawn as one factor |
Themes of experiments#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
a list of numbers |
Coarseness levels tried when grouping themes into broad themes |
|
|
|
|
Which similarities in the shuffled copies the chance cutoffs are read from |
|
|
|
number, more than 0, less than 1, or left empty |
How much closer than chance an experiment must be to the rest of its theme to be placed; blank places all |
|
|
|
whole number, from 1 to 100 |
How many shuffled copies of the experiments the chance cutoffs are drawn from |
|
|
|
whole number, from 0 to 8 |
What counts as one collection when weighing them: each root, or each folder just below a root |
|
|
|
whole number, at least 1 |
Fewest experiments a collection is weighed as if it had, so a tiny collection does not outweigh the rest |
|
|
|
number, from 0 to 1 |
How much a large collection outweighs a small one: 1 counts every experiment alike, 0 every collection alike |
|
|
|
number, more than 1 |
How much a theme’s recent share must grow, or shrink, to be called rising or falling |
|
|
|
whole number, at least 1 |
How many latest years a theme’s recent share is taken over |
|
|
|
number, more than 0, less than 1, or left empty |
How much closer than chance two experiments of different collections must be to be joined; blank joins every nearest neighbor |
|
|
not set |
number, more than 0, less than 1, or left empty |
How much closer than chance two experiments of one collection must be to be joined; blank joins every nearest neighbor |
|
|
|
|
Which of the drawn seeds is shown: the one most like the others, or the first |
|
|
|
a list of numbers |
Coarseness levels tried when drawing themes; the steadiest readable one is kept |
|
|
on |
on or off |
Whether a word the Dictionary does not know and its plural count as one word when both appear |
|
|
|
a list of words |
Kinds of Dictionary word that say how a file was made, left out of themes |
|
|
empty list |
a list of words |
Words left out of every theme |
|
|
|
whole number, from 0 to 1000 |
How many times the experiments are drawn again to see how steady each label is; 0 draws none |
|
|
|
|
How a theme’s words are ranked: by TF-IDF, or by the log-odds of the word inside the theme against the rest |
|
|
|
number, more than 0 |
How strongly the log-odds ranking leans on how common a word is overall |
|
|
|
number, from 0 to 1 |
Share of a Dictionary word’s use that must fall in a theme for it to lead the theme’s label |
|
|
|
whole number, at least 1 |
How many words a theme’s label shows |
|
|
not set |
whole number, at least 2, or left empty |
Most dimensions experiments’ words are summarized in before they are compared; blank grows them with the experiments |
|
|
|
number, more than 0, at most 1 |
A word in more than this share of experiments is background, not a theme word |
|
|
|
whole number, at least 1 |
Fewest experiments that must share a word the Dictionary knows for it to count |
|
|
|
whole number, at least 1 |
Fewest experiments of one collection that must share an unknown word for it to count |
|
|
|
whole number, at least 1 |
Fewest experiments with different words a theme needs; experiments that read the same words are not a theme on their own |
|
|
|
whole number, at least 1 |
Fewest experiments a theme needs; smaller groups are listed as not placed |
|
|
not set |
whole number, at least 1, or left empty |
How many closest experiments each one is joined to; blank chooses from the number of experiments |
|
|
|
whole number, from 1 to 8, or left empty |
Which folder level below each root is a person’s, for who else; blank for nobody |
|
|
|
whole number, at least 1 |
Fewest experiments a person needs in a theme, or with a word, to count there |
|
|
|
whole number, from 0 to 8 |
How many coarseness levels past the last one tried are tried when the steadiest readable one is the last |
|
|
|
whole number, from 2 to 64 |
How many times the themes are drawn, to measure how steady each one is |
|
|
|
number, more than 0, at most 1 |
Most of a theme’s experiments a word may be in and still be one of its specific words |
|
|
|
number, more than 0, at most 1 |
Share of the experiments with a word that must be in the theme for it to be the theme’s |
|
|
|
number, at least 0 |
How much a word in a folder inside an experiment counts |
|
|
off |
on or off |
Whether a word written many times counts less than in proportion |
|
|
|
number, from 0 to 1 |
Most often two words may appear together and still be read as one replacing the other |
|
|
|
number, from 0 to 1 |
How alike the company two words keep must be to read one as replacing the other |
|
|
|
whole number, at least 1 |
How many distinctive words are kept for each theme |
|
|
|
number, at least 0 |
How much a word in an experiment folder’s own name counts |
|
|
on |
on or off |
Whether a word must appear in a folder name or an acquisition’s name to count |
|
|
off |
on or off |
Whether identical experiments, gathered into one, keep their number’s weight when themes are drawn |
|
|
|
whole number, at least 1 |
Fewest experiments a year needs to be described in the year profile |
The question sheet#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
whole number, from 0 to 10000 |
A pattern of made files seen in this many pairs or fewer is asked about as unusual |
|
|
|
whole number, from 0 to 10000 |
The most questions about unusual pairs on the sheet |
|
|
|
number, from 0 to 1 |
A pattern seen in this share of all pairs or fewer is unusual, where that is more than the floor |
|
|
|
number, more than 0, at most 1 |
The share of the rules’ readings the first questions about them are chosen to cover |
|
|
|
whole number, from 0 to 10000 |
The most questions about the rules’ readings asked first |
|
|
|
number, more than 0, at most 1 |
The share of the proposed made-from pairs the first pattern questions are chosen to cover |
|
|
|
whole number, from 0 to 10000 |
The most questions about patterns of made files asked first |
|
|
|
whole number, from 1 to 10000 |
The most questions on the question sheet |
|
|
|
whole number, from 0 to 10000 |
The most words standing between two meanings the sheet asks about |
The search box#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
whole number, from 0 to 200 |
How many words that come with a search are listed beside its results |
|
|
|
whole number, at least 1 |
How many matched results a word must be in before it is listed as coming with a search |
|
|
|
whole number, from 1 to 64 |
The shortest word a search finds at the start of a longer word, or inside a name |
Tables with the same columns#
Setting |
Default |
May be |
What it is for |
|
|---|---|---|---|---|
|
|
whole number, from 0 to 500 |
How many files each list on a table’s page names; the count beside the list is always the whole |
|
|
on |
on or off |
Whether tables are listed whose columns contain another table’s columns |
|
|
|
whole number, from 1 to 1000 |
Fewest columns a table needs to be listed as contained in another table |
|
|
|
whole number, from 1 to 1000 |
Fewest columns a table needs to be listed as having the same columns as other tables, used only when the rule for tables with the same columns is not in force |