CLI reference¶
Generated from the command-line help at build time.
mcaligner¶
Usage: mcaligner [OPTIONS] COMMAND [ARGS]...
LLM agents for word alignment of parallel texts.
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --version Show the version and exit. │
│ --install-completion Install completion for the current shell. │
│ --show-completion Show completion for the current shell, to copy │
│ it or customize the installation. │
│ --help Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Commands ───────────────────────────────────────────────────────────────────╮
│ evaluate Precision, recall and F1 over sure and over all links, and the │
│ alignment error rate (AER). │
│ lexicon Export the bilingual lexicon: lemma -> aligned lemmas with counts │
│ and parts of speech. │
│ rules Link near-certain word pairs without a model, or measure the │
│ rules' precision (--measure). │
│ run Align with rules, then aligner/scorer agents; prints tokens and │
│ estimated cost. │
│ settings Print the settings (defaults, or a file's values with defaults │
│ filled in) as YAML. │
│ mcp MCP server over the corpora and alignment sets, with alignment │
│ exploration tools. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner rules¶
Usage: mcaligner rules [OPTIONS]
Link near-certain word pairs without a model, or measure the rules' precision
(--measure).
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ * --source -s <str> [NAME=]PATH[:ro|rw[:CONF… │
│ CoNLL-U source (default │
│ ro). Repeatable. │
│ [required] │
│ * --alignment -a <str> NAME=SOURCE_A,SOURCE_B,P… │
│ alignment set (default │
│ rw). Repeatable. │
│ [required] │
│ --target <str> Writable alignment set to │
│ add the rules' links to. │
│ --measure <str> Instead of writing, │
│ measure each rule's │
│ precision against this │
│ set. │
│ --evidence <str> Alignment sets used as │
│ lexicon evidence (default │
│ all). Repeatable. │
│ --min-count <int> Linked occurrences needed │
│ by the lexicon rule. │
│ [default: 5] │
│ --min-share <float> Share of occurrences that │
│ must agree. │
│ [default: 0.95] │
│ --show-errors --no-show-errors With --measure, list the │
│ links not in the │
│ reference set. │
│ [default: no-show-errors] │
│ --help Show this message and │
│ exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner run¶
Usage: mcaligner run [OPTIONS]
Align with rules, then aligner/scorer agents; prints tokens and estimated
cost.
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ * --source -s <str> [NAME=]PATH[:ro|r… │
│ CoNLL-U source │
│ (default ro). │
│ Repeatable. │
│ [required] │
│ * --alignment -a <str> NAME=SOURCE_A,SOU… │
│ alignment set │
│ (default rw). │
│ Repeatable. │
│ [required] │
│ * --target <str> Writable alignment │
│ set to fill (an │
│ --alignment name). │
│ [required] │
│ --evidence <str> Alignment sets │
│ used as evidence │
│ (default all). │
│ Repeatable. │
│ --sentence <str> Only these pairs │
│ (sent_id). │
│ Repeatable. │
│ --max-sentences <int> Only the first N │
│ pairs (after │
│ --sentence). │
│ --settings <path> YAML settings file │
│ (see `settings`). │
│ --rounds <int> Maximum │
│ align/score │
│ rounds. │
│ --concurrency <int> Agent tasks at the │
│ same time. │
│ --batch-sentences <int> Sentence pairs per │
│ agent task. │
│ --threshold <float> Links below this │
│ confidence are │
│ revisited. │
│ --model <str> One model for │
│ every role │
│ (overrides the │
│ tiers). │
│ --strong-model <str> Strong tier model │
│ (aligner). │
│ --efficient-model <str> Efficient tier │
│ model (scorer). │
│ --light-model <str> Light tier model. │
│ --aligner-model <str> Model for the │
│ aligner (overrides │
│ its tier). │
│ --scorer-model <str> Model for the │
│ scorer (overrides │
│ its tier). │
│ --strong-thinking <str> Reasoning effort │
│ for the strong │
│ tier: off, on, │
│ minimal..xhigh. │
│ --efficient-think… <str> Reasoning effort │
│ for the efficient │
│ tier. │
│ --scorer --no-scorer Run the scorer │
│ (default from │
│ settings: on). │
│ --guidelines <str> Annotation │
│ guidelines for the │
│ agents: a bundled │
│ name (grc-lat, the │
│ default), a file │
│ path, or none. │
│ --columns <str> Columns shown for │
│ each word after id │
│ and form, e.g. │
│ lemma,upos,head,d… │
│ none = forms only. │
│ --tools <str> Read tools the │
│ agents may call: │
│ all, none, or a │
│ comma-separated │
│ list. │
│ --link-punctuation --no-link-punctua… Allow links to │
│ punctuation │
│ (default: dropped │
│ in code). │
│ --report <path> Write the run │
│ report (JSON) │
│ here. │
│ --help Show this message │
│ and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner evaluate¶
Usage: mcaligner evaluate [OPTIONS] {predicted} {gold}
Precision, recall and F1 over sure and over all links, and the alignment error
rate (AER).
╭─ Arguments ──────────────────────────────────────────────────────────────────╮
│ * predicted <path> Predicted alignment set (JSONL). [required] │
│ * gold <path> Gold alignment set (JSONL); only its pairs are │
│ scored. │
│ [required] │
╰──────────────────────────────────────────────────────────────────────────────╯
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --annotated-only --no-annotated-only Score only pairs with at │
│ least one predicted │
│ link. │
│ [default: │
│ no-annotated-only] │
│ --min-confidence <float> Ignore predicted links │
│ below this confidence. │
│ [default: 0.0] │
│ --per-sentence --no-per-sentence Print a row per sentence │
│ pair. │
│ [default: │
│ no-per-sentence] │
│ --json Print the full report as │
│ JSON. │
│ --help Show this message and │
│ exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner lexicon¶
Usage: mcaligner lexicon [OPTIONS]
Export the bilingual lexicon: lemma -> aligned lemmas with counts and parts of
speech.
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ * --source -s <str> [NAME=]PATH[:ro|rw[:CONFIDENCE]] │
│ CoNLL-U source (default ro). │
│ Repeatable. │
│ [required] │
│ * --alignment -a <str> NAME=SOURCE_A,SOURCE_B,PATH[:ro|rw[:C… │
│ alignment set (default rw). │
│ Repeatable. │
│ [required] │
│ --output -o <path> YAML (or .json) file; default stdout. │
│ --use <str> Alignment sets to count (default all). │
│ Repeatable. │
│ --min-confidence <float> Only links at least this confident │
│ count. │
│ [default: 0.8] │
│ --min-count <int> Keep translations seen at least this │
│ often. │
│ [default: 1] │
│ --top <int> Keep at most this many translations │
│ per lemma. │
│ --help Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner settings¶
Usage: mcaligner settings [OPTIONS]
Print the settings (defaults, or a file's values with defaults filled in) as
YAML.
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --settings <path> YAML settings file to check and show. │
│ --help Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner mcp serve¶
Usage: mcaligner mcp serve [OPTIONS]
Serve sources and alignment sets over MCP (stdio for Claude Desktop, or
streamable HTTP).
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ * --source -s <str> [NAME=]PATH[:ro|rw[:CONFID… │
│ Repeat for several sources. │
│ Default mode is ro. │
│ [required] │
│ --transport <stdio|streamable-http> MCP transport. │
│ [default: stdio] │
│ --host <str> Host for streamable-http. │
│ [default: 127.0.0.1] │
│ --port <int> Port for streamable-http. │
│ [default: 8000] │
│ --read-only Register only the read-only │
│ tools. │
│ --root-key <str> MISC attribute holding the │
│ word's root. │
│ [default: Root] │
│ --name <str> Server name shown to MCP │
│ clients. │
│ [default: mcaligner] │
│ --alignment -a <str> NAME=SOURCE_A,SOURCE_B,PAT… │
│ word alignments between two │
│ sources. Repeatable. │
│ --corpus <str> [NAME=]PATH[,PATH...]: │
│ unannotated text (one │
│ sentence or paragraph per │
│ line) that agents can │
│ search for how a word is │
│ used. Repeatable. │
│ --help Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯
mcaligner mcp config¶
Usage: mcaligner mcp config [OPTIONS]
Print the JSON to add to Claude Desktop's claude_desktop_config.json.
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ * --source -s <str> [NAME=]PATH[:ro|rw[:CONFIDENCE]]. Repeat for │
│ several sources. Default mode is ro. │
│ [required] │
│ --read-only Register only the read-only tools. │
│ --root-key <str> MISC attribute holding the word's root. │
│ [default: Root] │
│ --name <str> Server name shown to MCP clients. │
│ [default: mcaligner] │
│ --alignment -a <str> NAME=SOURCE_A,SOURCE_B,PATH[:ro|rw[:CONFIDEN… │
│ word alignments between two sources. │
│ Repeatable. │
│ --corpus <str> [NAME=]PATH[,PATH...]: unannotated text (one │
│ sentence or paragraph per line) that agents │
│ can search for how a word is used. │
│ Repeatable. │
│ --help Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯