getClosingReport
ChatGPTThe closing report of a finished subproject (when closing_report_api is not null): verified results per success criterion, as sections and as a criteria table. Best first source for the main results. If the subproject was reopened later (closing_report_addenda > 0), "addenda" tell why and what changed, and "criteria" show the current state; the sections above the first addendum are the unchanged original report. If addenda is not empty, criteria and the addenda give the current state; the original report above the first addendum is a historical record.
getCodeFile
ChatGPTThe content of one text file of a code version (script or output), in parts of at most 90,000 bytes. When truncated is true and next_offset is not null, ask again with offset=next_offset for the next part. The file is returned as text and never executed.
getDerivation
ChatGPTA derivation in full (Markdown with LaTeX), the current version or any earlier one, with its key results, definitions, inputs, checks and reviews. Use section to get one section only. version is an integer (1, 2, ...); without it the current version is returned. inputs lists the results it uses, with their source and version.
getDocument
ChatGPTOne document of a reference subproject in full (Markdown with LaTeX), or one section of it, with its citation data. For example the rules of the description language Ilanguage 1.b: document ilanguage-1b of subproject 00-ilanguage. The version of a document is a string such as "1.b". Without section the whole document is returned; section must be an exact title from section_titles.
getEquation
ChatGPTOne equation by its tag (for example 15.7): its LaTeX in every version that has it, its section and link, the packages that use it, and the texts that cite it. In a reference subproject the tags are 0.k (for example 0.12 in 00-ilanguage). Tags are unique only within one subproject: resolve a qualified citation such as "(17.7) of 03-ilang-space/17-motion" in the subproject it names, at the version it names.
getGraph
ChatGPTThe dependency graph: every package with status, current version and regeneration counters, and every edge with the results (tags) it uses. cross_edges are results taken from packages of other subprojects ("<subproject>/<package>"), with the version used.
getPackage
ChatGPTOne package: goal, result, open issues, LaTeX of the equations cited in the result, versions with reviews, uses and used-by, citation data, and links to the question, derivation, reviews and code. uses_external lists the results taken from other subprojects, with the version used; versions lists every derivation version with its reviews.
getQuestion
ChatGPTThe question of a package in full (Markdown with LaTeX): goal, inputs quoted with their tags, assumptions, scope, expected result. The question states the goal and quotes the inputs with their tags and versions.
getResults
ChatGPTThe Result and Open issues sections of the current derivation of every package, with status and latest review verdict. The status "verified" means accepted by the research workflow after AI cross-checks, not independent human peer review.
getReview
ChatGPTOne review in full (Markdown with LaTeX) with its verdict and its list of issues. The verdict applies to version_reviewed only.
getSubproject
ChatGPTOverview of one subproject. Research: base problem sections (research question, model, assumptions, notation, success criteria), success criteria, and all packages with status, version, date, dependencies and addressed criteria. Reference: its documents with version, date, how to cite, section titles and equation tags. group names its research programme and stage, with the stages it builds on and is continued in.
listCode
ChatGPTThe code of a package: every code version with its scripts and outputs, their roles, sizes and links. Code versions are strings such as "v1".
listReviews
ChatGPTThe reviews of a package: kind, round, reviewed version, verdict, date and issues. version_reviewed is the derivation version a review assessed: its verdict applies to that version only. Reviews are by AI models of other providers, not human peer review.
listSubprojects
ChatGPTList the published subprojects. kind research: summary, state and package counts by status; kind reference: the reference documents, such as the definition of the description language Ilanguage. "groups" lists the research programmes: subprojects that build on each other as stages. "about" gives the author, the license and the Human–AI note (who contributed what). kind=reference (00-ilanguage) holds fixed reference documents, read with getDocument; it has no packages, graph, results or closing report.
listTopics
ChatGPTList the research topics: what the site studies. Start here. Each topic has a title, a description and what belongs to it: its research programmes with their stages (subprojects that build on each other) and its subprojects outside a programme, each with title, summary, state and links. "reference" lists the reference subprojects (the definition of the description language Ilanguage), which belong to no topic. The starting point for a general question. The title and description of a topic state a research goal, not a proven result.
search
ChatGPTFull-text search in base problems, closing reports, questions, derivations, reviews and reference documents. Every word of q must occur in the same paragraph. Returns snippets with links; paginate with limit and offset. Returns snippets only: read the whole text of a hit with the tool for its document (getDocument, getDerivation, getQuestion, getReview, getClosingReport or getSubproject), using its subproject, package, document and version. total is the number of hits; page with offset.
