Pandoc, YAML, JSON, CSV, Sqlite

README.md

Pandoc, YAML, JSON, CSV, Sqlite

This gists shows a little idea: use Pandoc to format data into HTML, or any other output format supported by Pandoc. This uses the ability of Pandoc to uses templates, and to receive data as YAML either from a separate file, or within a metadata block contained in the document itself..

It also shows a few scripts to use other data formats than YAML and extract metadata blocks as JSON. Note that multiple metadata blocks can be present within a single document.

Self-contained Pandoc document

The document.md file contains data in its metadata YAML block and is its own template:

$ pandoc --template document.md document.md
---
title: The document title
host:
- name: host-1
  ip: 172.17.0.1
- name: host-2
  ip: 172.17.0.2
---

Some hosts:

- host-1: 172.17.0.1
- host-2: 172.17.0.2

The result is a normal Markdown document that can be processed with Pandoc as usual, e.g. to produce an HTML page.

Separate data

The data can also come from an external file (here, hosts.yml):

$ pandoc --template document.md hosts.yml document.md
---
title: The document title
host:
- name: host-1
  ip: 172.17.0.1
- name: host-2
  ip: 172.17.0.2
---

Some hosts:

- host-1: 172.17.0.1
- host-2: 172.17.0.2
- host-3: 172.17.0.3

Note: here the hosts key in the metadata provided by document.md is not used because Pandoc is left-biased when merging multiple metadata blocks.

Other formats

Several scripts are provided to use other data format. For instance:

• From hosts.json (one object per line) to CSV,
• From CSV to a Sqlite database,
• From a Sqlite database to JSON again (a list of objects),
• From a Sqlite database to YAML (using the previous script).

This shows that the data rendered by Pandoc can come from Sqlite with minimal efforts.

Extract metadata block

It is also possible to extract the data "stored" within a metadata block by using a dedicated template (here, metadata.tpl):

$ pandoc --template metadata.tpl document.md
{"title":"The document title","host":[{"ip":"172.17.0.1","name":"host-1"},{"ip":"172.17.0.2","name":"host-2"}]}

csv-to-sqlite.sh

#! /usr/bin/env bash

sqlite3 hosts.db -cmd '.mode csv' '.import hosts.csv hosts'

# Demonstrate the import succeeded
sqlite3 hosts.db '.schema hosts'
sqlite3 hosts.db 'select * from hosts'

document.md

title	host
The document title	name ip host-1 172.17.0.1 name ip host-2 172.17.0.2

Some hosts:

$for(host)$

• $host.name$: $host.ip$ $endfor$

hosts.json

{ "name":"host-1", "ip":"172.17.0.1" }
{ "name":"host-2", "ip":"172.17.0.2" }
{ "name":"host-3", "ip":"172.17.0.3" }

json-to-csv.sh

#! /usr/bin/env bash

# Construct the CSV header
head -1 hosts.json | jq -r 'keys | @csv' > hosts.csv

# Extract rows
cat hosts.json | jq -r '[.[]] | @csv' >> hosts.csv

metadata.tpl

$meta-json$

sqlite-to-json.sh

#! /usr/bin/env bash

sqlite3 hosts.db \
  "select json_group_array(json_object('ip', ip, 'name', name)) as host
  from (select * from hosts)" | jq .

sqlite-to-yaml.sh

#! /usr/bin/env bash

echo '---'
echo host:
./sqlite-to-json.sh | jq -r -f to-yaml.jq
echo '---'

to-yaml.jq

def y:
  (objects | to_entries[] | (.value | type) as $type |
    if $type == "array" then
      "\(.key):", (.value | y)
    elif $type == "object" then
      "\(.key):", "  \(.value | y)"
    else
      "\(.key): \(.value)"
    end
  )
  // (arrays | select(length > 0)[] | [y] |
    "  - \(.[0])", "    \(.[1:][])"
  )
  // .
  ;
y