CLI
ftm-analyze exposes a Typer command-line interface. Run ftm-analyze --help to see the same content rendered in your terminal.
ftm-analyze
Top-level entry point. Without a subcommand it prints help.
| Option | Description |
|---|---|
--version / --no-version |
Print the installed version and exit. |
--settings / --no-settings |
Print the current Settings (resolved from environment variables and .env) and exit. |
--install-completion |
Install shell completion for the current shell. |
--show-completion |
Print the shell completion script. |
--help |
Show help and exit. |
Commands
download-spacy– fetch the spaCy language models referenced by the current settings.analyze– analyze a stream of FollowTheMoney entities.analyze-text– analyze a single text string (debug helper).
download-spacy
Download the spaCy models referenced by FTM_ANALYZE_SPACY_MODELS_*. The download targets are determined from Settings().spacy_models, so override the language-to-model mapping via environment variables before running this command if you want non-default models.
| Option | Description |
|---|---|
--help |
Show help and exit. |
analyze
Analyze a stream of FollowTheMoney entities. Reads entities from -i and writes annotated entities + mention fragments to -o. Both URIs go through anystore, so file paths, s3://, http(s)://, and database URIs (e.g. postgresql:///ftm) all work; - means stdin / stdout.
| Option | Default | Description |
|---|---|---|
-i TEXT |
- |
Input entities URI (file, http, s3, …). |
-o TEXT |
- |
Output entities URI (file, http, s3, …). |
--resolve-mentions / --no-resolve-mentions |
from FTM_ANALYZE_RESOLVE_MENTIONS |
Resolve known mentions against juditha. |
--annotate / --no-annotate |
from FTM_ANALYZE_ANNOTATE |
Annotate extracted patterns, names, and mentions in text properties (see annotate). |
--validate-names / --no-validate-names |
from FTM_ANALYZE_VALIDATE_NAMES |
Drop mentions that juditha cannot validate against known name tokens. |
--refine-mentions / --no-refine-mentions |
from FTM_ANALYZE_REFINE_MENTIONS |
Refine NER tag for each mention using extra heuristics + juditha.lookup. |
--refine-locations / --no-refine-locations |
from FTM_ANALYZE_REFINE_LOCATIONS |
Canonicalize location mentions via geonames_tagger. |
--overwrite-lang / --no-overwrite-lang |
from FTM_ANALYZE_OVERWRITE_LANG |
Ignore any pre-existing language property and rewrite detectedLanguage. |
--help |
Show help and exit. |
Example
ftm-analyze analyze \
-i s3://data/entities.ftm.json \
-o postgresql:///ftm \
--resolve-mentions \
--annotate
analyze-text
Analyze a single text string. The input is wrapped in a synthetic PlainText entity (bodyText), so this is mainly a debugging shortcut – use analyze for real ingestion runs.
| Option | Default | Description |
|---|---|---|
-i TEXT |
- |
Input text URI (file, http, s3, … – or - for stdin). |
-o TEXT |
- |
Output entities URI. |
--resolve-mentions / --no-resolve-mentions |
from settings | Resolve known mentions against juditha. |
--annotate / --no-annotate |
from settings | Annotate extracted patterns, names, and mentions. |
--validate-names / --no-validate-names |
from settings | Drop mentions that juditha cannot validate. |
--refine-mentions / --no-refine-mentions |
from settings | Refine NER tag for each mention. |
--refine-locations / --no-refine-locations |
from settings | Canonicalize location mentions. |
--help |
Show help and exit. |