SPLime Public Hub
Discover useful Python Objects.
Browse verified, versioned Functions, Pipelines, and Library Adapters. Inspect the contract, pin an exact release, and run it with the SPLime framework.
Public catalog
Verified releases
find_rules
Search numeric classification rules with a dataset and target column. Skip unsupported columns and return their names and reasons in unusable_columns, alongside the model, predictions and holdout metrics.
hn_topic_digest
Fetch real HN stories, normalize them, apply the required immutable embedded trained model, count observed attention and render notebook-safe HTML. Training is a separate operation. Published implementation executes locally in embedded mode.
render_digest
Render observed HN topics, tags, attention summaries, collection coverage and real story links as an escaped notebook-safe HTML fragment with scoped CSS; optional standalone document.
case5_first_pd1_model
Retrieve the fitted first_pd1 JSON model, exact rules and input schema embedded in this immutable release.
case5_first_pd1_predict
Apply the fitted first_pd1 rules embedded in this release. Returns application IDs, predicted classes, probabilities and segments. No caller-side model file or retraining is required.
evaluate_topics_temporally
Split HN stories chronologically, fit vocabulary and topics only on earlier titles, then inspect later-title coverage and reconstruction. No supervised accuracy or future-popularity forecast is claimed.
fetch_snapshot
Collect real HN stories from date-filtered Algolia search or current Firebase feeds. Explicit time range, overlapping feeds, request/item/time budgets and collection timestamp; no substitute data. Returns normalized rows and coverage diagnostics.
normalize_stories
Normalize HN records or a snapshot, preserve missing popularity values, merge overlapping feed membership and deduplicate story IDs. Returns acquisition provenance and quality counts.
summarize_topics
Measure observed topic counts, shares, scores and comments separately from topic assignment; compare equal-duration windows under one fixed model and report overlapping feed coverage.
analyze_titles
Apply explicit topic rules and compare document shares in adjacent equal-duration windows. Counts describe the supplied snapshot, not all Hacker News activity.
hn_topic_model
Return the immutable fitted JSON topic model included as a native Scalar dependency in this signed release. No model file, caller model input or retraining is required.
predict_topics
Apply a supplied fitted HN topic model without fitting vocabulary or components. Returns topic assignments, automatic keyword tags, unknown-topic diagnostics and unchanged observed popularity values.
tag_titles
Assign HN titles to topics with the immutable trained model embedded as a required native graph dependency. Automatic names use learned top terms; no caller model file or retraining. Returns hn-topics-v1 predictions.
train_topic_model
Fit a simple TF-IDF and nonnegative matrix factorization topic model on supplied training titles only. Returns portable JSON vocabulary, IDF weights, learned components, automatic top-term names and a content identity; no rule-generated labels.
numeric_rule_experiment
Reproduce a shallow numeric-rule experiment: reserve an external IID holdout, search and fit on development data using exact public stage versions, evaluate the fixed model, and report holdout errors. Caller supplies a numeric pandas DataFrame and target_col; code executes locally.
numeric_rule_partition
Validate a numeric pandas DataFrame and target column, then reserve a fixed stratified IID external holdout before fitting transformations or selecting rules. Returns original training and holdout row positions.
rule_complete_report
Add application-keyed predictions and evaluation partition labels to a completed Case5 training report.
rule_evaluate
Evaluate a fixed fitted model on development and independent holdout without changing the selected rules.
rule_finalize
Produce a JSON model, exact rules, leaderboard, metrics and audit; omit the intermediate training table.
rule_fit
Fit the selected configuration using development data only.
rule_inspect_model
Validate and inspect a caller-supplied JSON rule model, classes, rules and feature schema.
rule_monitor
Monitor coverage and missing values. Pass optional target_col for labelled quality metrics.
rule_predict
Score a caller-supplied validated rule model without retraining or requiring a target.
rule_prepare
Validate a prepared Case5 table and apply the chronological grouped search preset.