2. RISE’s reporting system

2.1. Description

Note

Alongside the LaTeX/PDF system below, webreport(m) builds a self-contained interactive HTML report of a model in one command – equations, calibration and steady state per regime, and an interactive IRF explorer – with no LaTeX, no server and no external assets: the file opens offline in any modern browser and can be emailed as-is. With a model carrying several parameterizations, one report per parameterization is written (_p1, _p2, … suffixes; pages are never flattened into regime columns), or select one with webreport(m, 'parameterization', k).

RISE ships its own LaTeX-driven reporting system. The user-facing class is rnotes: a handle class that accumulates sections, text, figures, tables and equations, and renders them to PDF via pdflatex. A legacy class rprt is kept as a thin subclass of rnotes to preserve the historical positional constructor; new code should use rnotes directly.

The system is PDF-only. PDF compilation requires a working LaTeX distribution on PATH (MiKTeX on Windows, TeX Live / MacTeX on Linux and macOS). rise_startup probes for pdflatex and warns if it is missing.

2.2. Quick start

A minimal report:

r = rnotes(title="My first report", author="Ola Normann");
r.section("Introduction")
r.text("This is the body of the introduction.")
publish(r, "report.pdf")

r is a handle, so all subsequent r.section(...), r.figure(...), r.table(...) calls accumulate into the same report. publish is the final step and writes the .tex and .pdf next to the working directory.

Multiple authors, with optional email and institution:

authors = ["Alice Adams",   "alice@example.org", "University of A";
           "Bob Brown",     "bob@example.org",   "University of B";
           "Carol Clarke",  "",                  "University of C"];
r = rnotes(title="Switching DSGE: Policy or Volatility?", author=authors);

The author argument accepts a string scalar (one author, name only), a column string vector (multiple authors, names only), or an N x {1,2,3} string array where the extra columns are email and/or institution. Email is detected automatically by the presence of @.

2.3. Structuring the document

Sectioning commands mirror LaTeX:

r.chapter("A chapter")          % only with DocumentClass="report" or "book"
r.section("A section")
r.subsection("A subsection")
r.subsubsection("A subsubsection")
r.paragraph("A paragraph heading")
r.subparagraph("A sub-paragraph heading")

Each can be made un-numbered:

r.section("Acknowledgements", Numbered=false)

Sectioning commands register an internal label automatically; you can also pass an explicit one for cross-referencing:

r.section("Methods", Label="sec:methods")

A table of contents is requested at publish time:

publish(r, "report.pdf", TOC=true)

2.4. Adding content

Text, lists and quotes:

r.text("This paper studies the effect of policy on output.")
r.itemize({"First point", "Second point", "Third point"})
r.enumerate({"Step one", "Step two", "Step three"})
r.description({
    "DSGE", "Dynamic Stochastic General Equilibrium"
    "RBC",  "Real Business Cycle"
})
r.quote("A pithy quotation.")
r.verbatim({"function y = f(x)", "  y = x.^2;", "end"})
r.footnote("This is a footnote attached to the previous paragraph.")
r.url("https://www.example.org")

Equations are added by passing a LaTeX expression. The class auto-wraps plain expressions in an equation environment; if the expression already contains an environment (\begin{align}, \begin{gather}, etc.) it is left untouched.

r.equation("y_t = \rho y_{t-1} + \varepsilon_t")
r.equation("E_{t} \pi_{t+1} = \beta \pi_t + \kappa x_t", ...
           Numbered=false, Label="eq:nkp")

Figures can be added from an existing handle:

fh = figure;
x = 1:100;
plot(x, log(x), "linewidth", 2);
r.figure(fh, Caption="Log of x for x = 1..100")
close(fh)

Or from a file on disk:

r.figure("path/to/saved.png", Caption="An imported figure")

For multi-panel figures, create_figure builds the figure from a specification of subplots:

subplots = {
    {ts("2000Q1", cumsum(randn(60,1))), "caption", "Series A"}
    {ts("2000Q1", cumsum(randn(60,1))), "caption", "Series B", ...
     "highlighted_dates", {date2serial({"2008Q4","2009Q4"})}}
};
r.create_figure(subplots, Caption="Two random walks")

Tables accept numeric arrays, cell arrays, or MATLAB table objects:

data    = randn(4, 3);
headers = ["A", "B", "C"];
rows    = ["Row 1"; "Row 2"; "Row 3"; "Row 4"];

r.table(data, ...
        Headers   = headers, ...
        RowNames  = rows, ...
        Caption   = "Sample table", ...
        Precision = 3)

For tables with more than MaxRows rows (default 10) table emits a longtable so the table breaks across pages naturally.

Conditional formatting applies a colour or weight to cells passing a test:

fmt = struct("test", @(x) x > 0.5, "action", {"red","bold"});
r.table(data, Headers=headers, FormatFuncs=fmt)

2.5. Page control

The page-control primitives mirror LaTeX:

r.newpage()           % \\newpage  - force a new page now
r.pagebreak()         % \\pagebreak - suggest a break here
r.clearpage()         % \\clearpage - flush pending floats
r.cleardoublepage()   % \\cleardoublepage - next page is odd-numbered

2.6. Publishing

publish compiles the accumulated body to PDF. The first argument after obj is the savefile name; it must be a filename only, with no path component. All formatting choices live on publish as name-value arguments:

publish(r, "report.pdf", ...
        DocumentClass = "report", ...
        PaperSize     = "a4paper", ...
        PointSize     = "11pt", ...
        Orientation   = "landscape", ...
        TOC           = true, ...
        NumPasses     = 2)

DocumentClass accepts article, report, book, letter, proc and minimal. NumPasses controls the number of pdflatex runs, which determines whether cross-references and the TOC resolve on the same compile (default 2). Packages and PreambleAdditions are escape hatches for advanced uses; the default preamble already loads graphicx, amsmath, amssymb, hyperref, xcolor, longtable, booktabs, listings and others.

2.7. Legacy `@rprt` compatibility

rprt is a thin subclass of rnotes that keeps the historical positional constructor working:

xrep = rprt("Title", "Ola Normann", "orientation", "landscape");

The legacy author shapes (struct array with Name / Email / Institution fields, cell of structs) and the legacy parameter names on individual methods (numbering, format_funcs, maxRows, precision, and the term / description struct form on description) are translated to their rnotes equivalents internally. The legacy method names plain, rise_file and rise_model_parameterization forward to text, dsge and addParameterTable respectively.

New code should not use rprt — the subclass exists only so that existing example scripts and external user code keep working without modification.