Use Task.async_stream for parallel org->md export. Configurable via EXPORT_CONCURRENCY env var or :export_concurrency config.
160 lines
4.8 KiB
Elixir
160 lines
4.8 KiB
Elixir
defmodule OrgGarden.Export do
|
|
@moduledoc """
|
|
Org-to-Markdown export via Emacs batch + ox-hugo.
|
|
|
|
Provides both single-file and batch export, plus a helper to compute
|
|
the expected `.md` output path for a given `.org` source file.
|
|
"""
|
|
|
|
require Logger
|
|
|
|
@doc """
|
|
Export a single `.org` file to Markdown via `emacs --batch` + ox-hugo.
|
|
|
|
Returns `{:ok, exit_code}` with the emacs exit code (0 = success),
|
|
or `{:error, reason}` if the command could not be executed.
|
|
"""
|
|
@spec export_file(String.t(), String.t(), String.t()) :: {:ok, non_neg_integer()} | {:error, term()}
|
|
def export_file(orgfile, notes_dir, output_dir) do
|
|
OrgGarden.Telemetry.span_export(orgfile, fn ->
|
|
do_export_file(orgfile, notes_dir, output_dir)
|
|
end)
|
|
end
|
|
|
|
defp do_export_file(orgfile, notes_dir, output_dir) do
|
|
section =
|
|
orgfile
|
|
|> Path.dirname()
|
|
|> Path.relative_to(notes_dir)
|
|
|
|
# ox-hugo requires static/ to exist for image asset copying
|
|
File.mkdir_p!(Path.join(output_dir, "static"))
|
|
|
|
{output, exit_code} =
|
|
System.cmd(
|
|
"emacs",
|
|
[
|
|
"--batch",
|
|
"--eval", "(require 'ox-hugo)",
|
|
"--eval", """
|
|
(org-cite-register-processor 'passthrough
|
|
:export-citation
|
|
(lambda (citation _style _backend _info)
|
|
(let ((keys (mapcar (lambda (ref)
|
|
(concat "@" (org-element-property :key ref)))
|
|
(org-cite-get-references citation))))
|
|
(format "[cite:%s]" (string-join keys ";")))))
|
|
""",
|
|
"--eval", "(setq org-cite-export-processors '((t passthrough)))",
|
|
"--eval", ~s[(setq org-hugo-base-dir "#{output_dir}")],
|
|
"--eval", ~s[(setq org-hugo-default-section-directory "#{section}")],
|
|
"--visit", orgfile,
|
|
"--funcall", "org-hugo-export-to-md"
|
|
],
|
|
stderr_to_stdout: true
|
|
)
|
|
|
|
filtered =
|
|
output
|
|
|> String.split("\n")
|
|
|> Enum.reject(&String.match?(&1, ~r/^Loading|^ad-handle|^For information/))
|
|
|> Enum.join("\n")
|
|
|
|
if filtered != "", do: Logger.info("emacs: #{filtered}")
|
|
|
|
if exit_code == 0 do
|
|
{:ok, exit_code}
|
|
else
|
|
{:error, {:emacs_exit, exit_code, filtered}}
|
|
end
|
|
rescue
|
|
e -> {:error, e}
|
|
end
|
|
|
|
@default_max_concurrency 8
|
|
|
|
@doc """
|
|
Export all `.org` files found under `notes_dir`.
|
|
|
|
Exports files in parallel for improved performance. The concurrency level
|
|
can be configured via the `:export_concurrency` application config or
|
|
the `EXPORT_CONCURRENCY` environment variable. Defaults to #{@default_max_concurrency}.
|
|
|
|
Returns `{:ok, count}` where `count` is the number of successfully
|
|
exported files, or `{:error, failures}` if any files failed.
|
|
"""
|
|
@spec export_all(String.t(), String.t()) :: {:ok, non_neg_integer()} | {:error, list()}
|
|
def export_all(notes_dir, output_dir) do
|
|
org_files =
|
|
Path.join(notes_dir, "**/*.org")
|
|
|> Path.wildcard()
|
|
|
|
if org_files == [] do
|
|
Logger.warning("No .org files found in #{notes_dir}")
|
|
{:ok, 0}
|
|
else
|
|
max_concurrency = get_concurrency()
|
|
Logger.info("Exporting #{length(org_files)} org file(s) from #{notes_dir} (concurrency: #{max_concurrency})")
|
|
|
|
results =
|
|
org_files
|
|
|> Task.async_stream(
|
|
fn orgfile ->
|
|
Logger.info(" exporting: #{orgfile}")
|
|
{orgfile, export_file(orgfile, notes_dir, output_dir)}
|
|
end,
|
|
max_concurrency: max_concurrency,
|
|
timeout: :infinity,
|
|
ordered: false
|
|
)
|
|
|> Enum.map(fn {:ok, result} -> result end)
|
|
|
|
failures =
|
|
Enum.filter(results, fn
|
|
{_, {:ok, _}} -> false
|
|
{_, {:error, _}} -> true
|
|
end)
|
|
|
|
if failures == [] do
|
|
{:ok, length(results)}
|
|
else
|
|
{:error, failures}
|
|
end
|
|
end
|
|
end
|
|
|
|
defp get_concurrency do
|
|
Application.get_env(:org_garden, :export_concurrency, @default_max_concurrency)
|
|
end
|
|
|
|
@doc """
|
|
Compute the expected `.md` path for a given `.org` file.
|
|
|
|
Uses the same section-mapping logic as ox-hugo: the relative directory
|
|
of the `.org` file within `notes_dir` becomes the section directory
|
|
under `content_dir`.
|
|
|
|
## Examples
|
|
|
|
iex> OrgGarden.Export.expected_md_path("/notes/bus/emt.org", "/notes", "/out/content")
|
|
"/out/content/bus/emt.md"
|
|
|
|
iex> OrgGarden.Export.expected_md_path("/notes/top-level.org", "/notes", "/out/content")
|
|
"/out/content/top-level.md"
|
|
"""
|
|
@spec expected_md_path(String.t(), String.t(), String.t()) :: String.t()
|
|
def expected_md_path(orgfile, notes_dir, content_dir) do
|
|
section =
|
|
orgfile
|
|
|> Path.dirname()
|
|
|> Path.relative_to(notes_dir)
|
|
|
|
basename = Path.basename(orgfile, ".org") <> ".md"
|
|
|
|
case section do
|
|
"." -> Path.join(content_dir, basename)
|
|
_ -> Path.join([content_dir, section, basename])
|
|
end
|
|
end
|
|
end
|