Tag generator for Rmarkdown files

rmdtags.py

#! /usr/bin/env python
# -*- coding: utf-8 -*-

from __future__ import print_function

help_text = """
Extracts tags from rmd files. Useful for the Tagbar plugin.

Usage:
Install Tagbar (http://majutsushi.github.io/tagbar/). Then, put this file
anywhere and add the following to your .vimrc:

let g:tagbar_type_rmd = {
          \   'ctagstype':'rmd'
          \ , 'kinds':['h:header', 'c:chunk', 'f:function', 'v:variable']
          \ , 'sro':'&&&'
          \ , 'kind2scope':{'h':'header', 'c':'chunk', 'f':'function', 'v':'variable'}
          \ , 'sort':0
          \ , 'ctagsbin':'/path/to/rmdtags.py'
          \ . 'ctagsargs':''
          \ }
"""

import sys
import re

if len(sys.argv) < 2:
    print(help_text)
    exit()

filename = sys.argv[1]

# filename = "test.Rmd"
print(filename)

re_header = re.compile(r"^(#+)([^{]+)(\{[^}]+\})?$")
re_chunk = re.compile(r"^```\{r ([^,]+)(,[^}]*)?\}$")
re_function = re.compile(r"^[\t ]*([^ ]+) *<- *function.*$")
re_variable1 = re.compile(r"^[\t ]*([^ ]+) *<-.*$")
re_variable2 = re.compile(r"^.*-> *(.+)$")

file_content = []
try:
    with open(filename, "r") as vim_buffer:
        file_content = vim_buffer.readlines()
except:
    exit()

headers = []
depth = [0]
inChunk = False
curChunk = None
for lnum, line in enumerate(file_content):
    newline = []
    tag = ""
    signature = ""
    options = ""
    # Headers
    if re_header.match(line) and not inChunk:
        level = len(re_header.match(line).group(1))
        tag = re_header.match(line).group(2).strip()
        newline.append(tag)
        newline.append(filename)
        newline.append('/^' + line.rstrip("\n") + '$/;"')
        newline.append("h")
        newline.append("line:" + str(lnum+1))
        # header
        while (level <= depth[len(depth) -1]):
            headers.pop()
            depth.pop()
        if len(headers) > 0:
            options = "header:" + "&&&".join(headers)
        headers.append(tag)
        depth.append(level)
        # signature
        if re_header.match(line).group(3):
            signature = "(" + re_header.match(line).group(3).strip("{}") + ")"
            options = options + "\tsignature:" + signature
        # Print
        newline.append(options)
        print("\t".join(newline))
    # Chunks
    if re_chunk.match(line):
        inChunk = True
        curChunk = re_chunk.match(line).group(1).strip()
        newline.append(re_chunk.match(line).group(1).strip())
        newline.append(filename)
        newline.append('/^' + line.rstrip("\n") + '$/;"')
        newline.append("c")
        newline.append("line:" + str(lnum+1))
        if len(headers) > 0:
            options = "header:" + "&&&".join(headers)
        if re_chunk.match(line).group(2):
            signature = "(" + re_chunk.match(line).group(2).lstrip(", ") + ")"
            options = options + "\tsignature:" + signature
        # Print
        newline.append(options)
        print("\t".join(newline))
    if line == "```\n":
        inChunk = False
        curChunk = None
    # Functions
    if re_function.match(line) and inChunk:
        newline.append(re_function.match(line).group(1).strip())
        newline.append(filename)
        newline.append('/^' + line.rstrip("\n") + '$/;"')
        newline.append("f")
        newline.append("line:" + str(lnum+1))
        print("\t".join(newline))
    elif (re_variable1.match(line) or re_variable2.match(line)) and inChunk:
        tag = re_variable1.match(line) or re_variable2.match(line)
        tag = tag.group(1).strip()
        newline.append(tag)
        newline.append(filename)
        newline.append('/^' + line.rstrip("\n") + '$/;"')
        newline.append("v")
        newline.append("line:" + str(lnum+1))
        print("\t".join(newline))

Hi Maxime, thanks for this tag generator. I found this gist through your blog. I've been using your script for the past 6 months, maybe. I would like to expand the list of tags and include inline chunks that are located in the main text as:

Our data have `r nrow(df)` observations...

I think the variable for finding inline chunks could be (but I may be wrong):

re_inlinechunk = re.compile(r"^`r ([^,]+)(,[^}]*)?`$")

But I have no idea how to allocate it in your code. Any idea? Thanks again!

Hi Karlo,

I'm not using nvim anymore (hence not relying on this script anymore), so I won't be able to test it through and through.

You should remove the beginning of line and end of line anchors in your regexp, also there is no need to check for chunk options, so this should suffice:

re_inlinechunk = re.compile(r"`r ([^`])+`")

As for implementing the tag, you should insert this block (or something close to that, I haven't tested it) probably just before the logic for matching headers (# Headers).
You can go further and add the logic to place the inline chunks inside the headers, in the same fashion as with the block chunks and other headers.

if re_inlinechunk.match(line) and not inChunk:
        newline.append(re_inlinechunk.match(line).group(1).strip())
        newline.append(filename)
        newline.append('/^' + line.rstrip("\n") + '$/;"')
        newline.append("c")
        newline.append("line:" + str(lnum+1))
        print("\t".join(newline))

This should add the inline chunks as regular chunks in tagbar.
If you want a different behavior, you have to change the "c" kind to a new letter (that is not already used by another kind of tag), and declare that tag kind in kinds and kind2scope

Hi Maxime,

I will try this and, hopefully, I will do my best to make it work.

Thank you very much for your time.