Enum SpamCommand 

pub(crate) enum SpamCommand {
    Fetch {
        account: Option<String>,
        verify: bool,
        json: bool,
        job: Option<String>,
    },
    Train {
        no_fetch: bool,
        no_replace: bool,
        errors: Option<u64>,
        scores: bool,
        settings: SettingsArgs,
        json: bool,
        job: Option<String>,
        by_itself: bool,
    },
    Eval {
        errors: u64,
        json: bool,
    },
    Status {
        json: bool,
    },
    DryRun {
        account: Option<String>,
        limit: usize,
        matrix: MatrixArgs,
        json: bool,
    },
    Features {
        domain: Option<String>,
        json: bool,
    },
    Review {
        limit: usize,
        json: bool,
    },
    Label {
        message: String,
        label: String,
        moving: bool,
        no_move: bool,
        json: bool,
    },
    Job {
        id: Option<String>,
        stop: bool,
        json: bool,
    },
    Import {
        file: Option<PathBuf>,
        source: Option<String>,
        remove: Option<String>,
    },
    Why {
        file: PathBuf,
    },
}

Variants

Fetch

Downloads the training corpus: every folder of every account (or one), what is new since the last time, each message checked as Sioul checks the mail it stores (its signatures, its sender), then the messages kept before without these checks; nothing changes on the server.

Fields

account: Option<String>

verify: bool

Only the checks of the messages kept without them (each fetched whole once, in memory only): nothing new is downloaded.

json: bool

What it did, as JSON.

job: Option<String>

Runs as the job of this id (sioul spam job): where it stands goes to its file.

Train

Brings the corpus up to date, then trains the filter; the table is replaced only if it sets aside no more ham than the one in place.

Fields

no_fetch: bool

Train on the corpus as it is, without downloading first.

no_replace: bool

A trial: train and say what it would do, changing nothing (the table, the one before, the language model, the last training’s summary).

errors: Option<u64>

Also list the test’s worst errors, this many of each kind, with the counts by account and folder and the numbers at other thresholds.

scores: bool

Each test message’s probability of spam, its label and whether that label is settled, in the JSON: numbers only, for a curve or another threshold.

settings: SettingsArgs

json: bool

What it did, as JSON.

job: Option<String>

Runs as the job of this id (sioul spam job): where it stands goes to its file.

by_itself: bool

Trains again by itself (sioul_learn::auto): without --job, its job started apart at once, as the window’s minute asks; as its job, at the lowest priority, stopped when the computer is unplugged or starts saving power.

Eval

Tests the table in place on the newest fifth of the corpus: its numbers, by account and folder, at other thresholds, and its worst errors.

Fields

errors: u64

The worst errors listed, this many of each kind: ham called spam (the surest first), then spam missed (the least sure first).

json: bool

Status

What the corpus holds, the outside material, the table, the last training, the jobs.

Fields

json: bool

DryRun

What your filter would do now with the mail in each inbox, as if it came now: counts per verdict and action, the messages it would move and those it would flag. Nothing is moved.

Fields

account: Option<String>

Only this account’s inbox.

limit: usize

Messages listed at most, in each list.

matrix: MatrixArgs

json: bool

Features

Each header feature’s mean over the corpus, by account and label: those that tell an account apart rather than spam from ham. Numbers only.

Fields

domain: Option<String>

Only the mail of this sender’s domain (or one under it).

json: bool

Review

What your own filter caught, as the Porch lists it: flagged, or moved into a Junk folder.

Fields

limit: usize

json: bool

Label

Says a message is spam or not, as the window’s Spam and Not spam do, in this device’s label log (every device and the next training read it). It only labels: --move also does the window’s act on the server.

Fields

message: String

The message: its file, or its Message-ID (mid:…).

label: String

spam, or ham (not spam).

moving: bool

The window’s act too: the keyword on the server, and the move (spam into the Junk folder; not spam out of it, back to the inbox).

no_move: bool

Only the label (the default).

json: bool

Job

A job run apart (an agent’s download or training): where it stands, then its result. Without an id, the jobs kept.

Fields

id: Option<String>

stop: bool

Asks it to stop at its next step.

json: bool

Import

Imports outside training material: one JSON object per line (gzipped or not), with label (“spam” or “ham”), date (RFC 3339 or Unix seconds), subject and text; from, headers, html and source if known. Kept apart from your own corpus, never shared; a source imported again replaces the one before. Or, with –remove, takes a source away.

Fields

file: Option<PathBuf>

The file to import.

source: Option<String>

The source’s name, when its lines do not say it (else the file’s name).

remove: Option<String>

Takes this source away: the next training learns without it.

Why

Why the filter judges a message as it does: the words and header facts that weighed (for you: it shows the message’s words).

Fields

file: PathBuf

Trait Implementations

impl FromArgMatches for SpamCommand

fn from_arg_matches(__clap_arg_matches: &ArgMatches) -> Result<Self, Error>

Instantiate Self from [ArgMatches], parsing the arguments as needed.

fn from_arg_matches_mut( __clap_arg_matches: &mut ArgMatches, ) -> Result<Self, Error>

Instantiate Self from [ArgMatches], parsing the arguments as needed.

fn update_from_arg_matches( &mut self, __clap_arg_matches: &ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

fn update_from_arg_matches_mut<'b>( &mut self, __clap_arg_matches: &mut ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

impl Subcommand for SpamCommand

fn augment_subcommands<'b>(__clap_app: Command) -> Command

Append to [Command] so it can instantiate Self via [FromArgMatches::from_arg_matches_mut]

fn augment_subcommands_for_update<'b>(__clap_app: Command) -> Command

Append to [Command] so it can instantiate self via [FromArgMatches::update_from_arg_matches_mut]

fn has_subcommand(__clap_name: &str) -> bool

Test whether Self can parse a specific subcommand