Enum JobType 

pub enum JobType {
    Email(EmailArgs),
    Deduplicate {
        source_bucket: Vec<String>,
        local_output: Option<PathBuf>,
        destination_bucket: Option<String>,
        concurrency: Option<usize>,
        transfers: Option<usize>,
        checkers: Option<usize>,
        tpslimit: Option<usize>,
        upload_only: bool,
        report_bucket: Option<String>,
        non_interactive: bool,
    },
    Rclone(RcloneArgs),
    Transform {

        input_file_type: Option<String>,
        video_quality: Option<String>,
        audio_quality: Option<String>,
        source: Option<String>,
        destination: Option<String>,
        local_output: Option<PathBuf>,
        concurrency: Option<usize>,
        download_transfers: Option<usize>,
        download_checkers: Option<usize>,
        upload_transfers: Option<usize>,
        max_pending_downloads: Option<usize>,
        tpslimit: Option<usize>,
        report_bucket: Option<String>,
        non_interactive: bool,
    
},
}

Variants

Email(EmailArgs)

Email jobs addressed through the keyring. See subcommands for the available actions.

Deduplicate

Recursively scans a bucket, always inflates every zip found (the zip container itself is never uploaded, only its inflated contents), content-hashes every file bucket-wide to keep one byte-identical copy of each, writes a human-readable merge report, and optionally uploads the result unencrypted to a (possibly different) bucket-config (ADR-0082). Every file is always processed and every zip is always expanded – there is no file-type or zip-expansion selection, and this job never offers encryption.

Fields

source_bucket: Vec<String>

Alias of a configured bucket-config to pull from. Repeat the flag for more than one (--source-bucket a --source-bucket b) – every listed bucket is downloaded into one shared local staging tree and deduplicated across the combined set, not per-bucket (ADR-0109). Interactively multi-selected from the configured bucket-configs when omitted and stdin is a terminal; at least one is required otherwise.

local_output: Option<PathBuf>

Local directory to stage and store output under. Defaults to a directory under the OS temp directory when omitted.

destination_bucket: Option<String>

Alias of a configured bucket-config to upload the deduped result to, once local processing is complete. Always uploaded unencrypted. Single bucket only (ADR-0109).

concurrency: Option<usize>

Maximum number of files to download/hash concurrently.

transfers: Option<usize>

rclone –transfers for the upload phase, now one bulk rclone copy rather than a per-file task queue (ADR-0136). Defaults to 8 when omitted. With --upload-only, this is the only transfer concurrency flag that has any effect.

checkers: Option<usize>

rclone –checkers for the upload phase. Defaults to 16 when omitted.

tpslimit: Option<usize>

rclone –tpslimit for the upload phase. Unset by default – no rate cap.

upload_only: bool

Resumes uploading an already-completed local deduplicate run instead of starting a new one: skips the bucket listing/download/hash/ placement phases entirely (and the source bucket credentials they’d otherwise need) and uploads straight from an existing --local-output’s result/ tree, picking up where a prior run’s upload phase left off via the same .staging/.uploaded index (ADR-0089). Requires a --local-output from a completed prior run (its .staging/.processed checkpoint must exist and its result/ must be non-empty) and a mandatory --remote-output.

report_bucket: Option<String>

Alias of a configured bucket-config this run’s report, the shared observability log, and a transcript of its printed output are uploaded to, always unencrypted, under a YYYY-MM-DD-job-name-{run-id}/ prefix (ADR-0100). Mandatory – interactively selected when omitted and stdin is a terminal; required otherwise.

non_interactive: bool

Suppress every wizard prompt (ADR-0114): each omitted input resolves via its non-interactive fallback (or errors if none exists) instead of asking, and the final “proceed?” confirmation is auto-answered yes.

Rclone(RcloneArgs)

Shells out to the external rclone binary for copy/delete operations against pigeon keyring BucketConfig aliases. rclone is always invoked against a config file pigeon itself generates from the keyring (ADR-0136, reversing ADR-0101/ADR-0110’s original raw-passthrough design) – never an externally-provisioned or user-general-purpose rclone.conf. See subcommands for the available actions. Requires rclone on PATH.

Transform

Transcodes image files (png/jpeg/heic) into full-size, maximum-quality .jpg, video files (mov/m4v/mp4, ADR-0122) into space-optimized H.265/HEVC .mp4 at a chosen --video-quality, audio files – an audio-only recording carrying an .mp4 extension (mp4-audio, ADR-0142) or m4a/mp3/aac/wav/opus (ADR-0153) – into Opus .opus at a chosen --audio-quality, or spreadsheets (numbers/xlsx/xls/csv, ADR-0149) into one .csv per sheet, or documents (ADR-0154) – doc/docx/pages/rtf/pptx/key/ epub into .pdf with pdf copied through, txt/md/markdown/mkd/rst/textclipping into .txt, and tsv into .csv (ADR-0112, pipeline reworked by ADR-0116): shells out to rclone to pull --source into a local staging tree, transcoding/copying each file locally and pushing it individually via rclone copyto the moment it’s ready, rather than waiting for the whole batch. A multi-sheet workbook is the one input kind that yields several outputs, each named {stem}-{hash}-{nn}of{tt}-{sheet}.csv so they stay recognizably related and a half-pushed workbook is detectable on a rerun (ADR-0149). --source/--destination are pigeon keyring BucketConfig aliases, optionally composed with a prefix within that bucket (alias:prefix/, ADR-0138); an omitted prefix defaults to --input-file-type’s own extension (source) or output extension (destination). Deduplication is never involved – every output file’s name is derived deterministically and is unique by construction; duplicate source content is deliberately retained at separate destination names. A single file’s transcode or push failure (e.g. a HEIC file the local ffmpeg build can’t decode, retried before giving up) is recorded and reported but never stops any other file from being processed and pushed. A rerun is resumable without any local state: “already done” is answered by listing the destination (ADR-0120 removed the local checkpoint). Requires rclone on PATH, plus ffmpeg/ffprobe for an image, video, or audio input kind, soffice (LibreOffice) for numbers and for doc/docx/pages/rtf/pptx/key, ebook-convert (Calibre) for epub, and pandoc for md/rst; xlsx/xls/csv/tsv/pdf/txt/textclipping need none of them (ADR-0149, ADR-0154).

Fields

input_file_type: Option<String>

Which input format this run transcodes: png, jpeg, heic, mov, m4v, mp4, mp4-audio, m4a, mp3, aac, wav, opus, numbers, xlsx, xls, csv, doc, docx, pages, rtf, pdf, epub, pptx, key, txt, md (or markdown/mkd), rst, textclipping, or tsv. Interactively selected when omitted and stdin is a terminal; required otherwise. mp4-audio is for audio-only recordings carrying an .mp4 extension (voice notes), transcoded to .opus; a file that actually has a video stream is refused rather than silently losing its video – embedded cover art does not count as video (ADR-0153). opus input is re-encoded rather than copied through, so a mixed library normalizes to one bitrate. The four spreadsheet kinds all output .csv, one file per non-empty sheet (ADR-0149); csv input is copied through byte-for-byte, as are pdf and txt (ADR-0154). key is sniffed before conversion and a file that is not an iWork container is refused – that extension is also conventional for a private key, and LibreOffice would otherwise render one into a readable PDF. md/rst are rendered to prose, which discards link URLs and all inline markup.

video_quality: Option<String>

Compression level for a true video input kind (mov/m4v/mp4 – not mp4-audio, which has no video stream): low, medium (default), high, or lossless – a small vetted libx265 CRF menu (ADR-0122). Ignored for every other input kind. Interactively selected when omitted, stdin is a terminal, and the input kind is video; defaults to medium otherwise.

audio_quality: Option<String>

Opus bitrate for an audio input kind (mp4-audio, m4a, mp3, aac, wav, opus): low (32k, speech), medium (48k), high (96k), or transparent (128k) – a small vetted menu (ADR-0153), never a raw bitrate. Ignored for every other input kind. Interactively selected when omitted, stdin is a terminal, and the input kind is audio; otherwise it defaults per kind: medium for mp4-audio, preserving ADR-0142’s choice for voice notes, and transparent for the rest, which are music or (for wav) lossless and would be badly served by a speech bitrate.

source: Option<String>

Where to pull from: a configured bucket-config alias, optionally composed with a prefix inside it – media is the whole bucket, media:mp4/ its mp4/ directory (ADR-0138). An omitted prefix defaults to --input-file-type’s own extension, so --source media --input-file-type mp4 means media:mp4/; an explicit prefix is always used verbatim. Interactively selected (bucket root, then the default prefix) when omitted and stdin is a terminal; required otherwise.

destination: Option<String>

Where to push to, same alias[:prefix/] form as --source (ADR-0138). An omitted prefix defaults to --input-file-type’s output extension (jpg for an image kind, mp4 for a video one, opus for every audio kind, csv for every spreadsheet kind), so one bucket can serve as both source and destination when the prefixes differ.

local_output: Option<PathBuf>

Local directory this run stages pulled source files, transcoded results, and bookkeeping under. Defaults to a directory under the OS temp directory when omitted.

concurrency: Option<usize>

Maximum number of files to convert concurrently (the local conversion phase only – rclone manages its own parallelism via --download-transfers/--download-checkers). Interactively prompted (a cores-based default) when omitted and stdin is a terminal; required otherwise.

download_transfers: Option<usize>

rclone –transfers for the download (pull) phase. Defaults to 8 when omitted.

download_checkers: Option<usize>

rclone –checkers for the download (pull) phase. Defaults to 16 when omitted.

upload_transfers: Option<usize>

Maximum number of concurrent per-file push (rclone copyto) operations. Defaults to 8 when omitted. Separate from --download-transfers to allow independent tuning of pull vs. push phases (ADR-0129).

max_pending_downloads: Option<usize>

Maximum number of downloaded but not-yet-processed files (pending transcode/push) before the pull phase is paused to allow the pipeline to catch up. Defaults to concurrency * 4 when omitted. Enables automatic disk-pressure backpressure without requiring manual tuning (ADR-0129).

tpslimit: Option<usize>

rclone –tpslimit, used by the pull phase only. Unset by default – no rate cap.

report_bucket: Option<String>

Alias of a configured bucket-config this run’s report, the shared observability log, and a transcript of its printed output are uploaded to, always unencrypted, under a YYYY-MM-DD-job-name-{run-id}/ prefix (ADR-0100). Mandatory – interactively selected when omitted and stdin is a terminal; required otherwise.

non_interactive: bool

Suppress every wizard prompt (ADR-0114): each omitted input resolves via its non-interactive fallback (or errors if none exists) instead of asking, and the final “proceed?” confirmation is auto-answered yes.

Trait Implementations

impl Debug for JobType

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl FromArgMatches for JobType

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 JobType

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