How to generate json using Jinja2 Template Engine

Question

I am trying to generate a json file from a .xlsx file.

so far I was able to get data from the file but I am not sure how to put them on json using jinja2. There is a problem in template structure. How can I solve this?

Output should be like this

  "Matches": {    
    "1": {
      "time": "19:00",
      "teams": "Team 1 - Team 2"
    },
    "2": {
      "time": "21:00",
      "teams": "Team 3 - Team 4"
    },
    ...
    ...
    ...
  }

My code is something like this. Obviously template section is wrong.

from openpyxl import load_workbook
from jinja2 import Template


start_coloumn_of_matches = 3
end_coloumn_of_matches = 20

wb = load_workbook(filename = 'myfile.xlsx')
sheet_ranges = wb['Sheet1']

keys = []
teams = []
times = []


for x in range(start_coloumn_of_matches, end_coloumn_of_matches + 1):
    team_column = 'A' + str(x)
    time_column = 'D' + str(x)
    teams.append(sheet_ranges[team_column].value)
    times.append(sheet_ranges[time_column].value)
    keys.append(x)



template = Template(''' 
"Matches": {
        {% for key in keys %}
      "{{key}}": 
        {% endfor %}
      {
        {% for team in teams %}
        "teams": "{{team}}",
        {% endfor %}
        {% for time in times %}
        "time": "{{time}}"
        {% endfor %}
        }

    }, 

    ''' )

print(template.render(teams = teams, times = times, keys = keys))

Answer 1

Manually constructing json runs the risk of accidentally producing an invalid json string. It's safer to use a tool to do this, and it keeps your templates less cluttered.

If you are using Jinja 2.9 or later, you can use the built-in tojson filter to convert Python objects* to json automatically.

>>> import pprint

>>> # Construct some test data
>>> matches = ['1', '2', '3']
>>> times = ['19:00', '21:00', '23:00']
>>> teams = ['Team 1 - Team 2', 'Team 3 - Team 4', 'Team 5 - Team 6']

>>> # Combine the data structures to match the required output
>>> match_data = [dict(zip(['time', 'team'], pair)) for pair in zip(times, teams)]
>>> combined = {x: y for x, y in zip(matches, match_data)}
>>> pprint.pprint(combined)
{'1': {'team': 'Team 1 - Team 2', 'time': '19:00'},
 '2': {'team': 'Team 3 - Team 4', 'time': '21:00'},
 '3': {'team': 'Team 5 - Team 6', 'time': '23:00'}}

>>> template = jinja2.Template("""{{ matches | tojson(indent=2) }}""")
>>> print(template.render(matches=combined))
{
  "1": {
    "team": "Team 1 - Team 2",
    "time": "19:00"
  },
  "2": {
    "team": "Team 3 - Team 4",
    "time": "21:00"
  },
  "3": {
    "team": "Team 5 - Team 6",
    "time": "23:00"
  }
}

For earlier Jinja releases, construct the json using the json package in Python's standard library, and then render the json in your template.

>>> import json    
>>> # Serialise our object as json; setting the indent argument gives
>>> # the pretty printed format that we want.
>>> jdata = json.dumps(combined,  indent=2) 
>>> print(jdata)
{
  "1": {
    "time": "19:00",
    "team": "Team 1 - Team 2"
  },
  "2": {
    "time": "21:00",
    "team": "Team 3 - Team 4"
  },
  "3": {
    "time": "23:00",
    "team": "Team 5 - Team 6"
  }
}

>>> # Pass the json to the template for rendering.
>>> template = jinja2.Template("""{{ matches }}""")
>>> print(template.render(matches=jdata))
{
  "1": {
    "time": "19:00",
    "team": "Team 1 - Team 2"
  },
  "2": {
    "time": "21:00",
    "team": "Team 3 - Team 4"
  },
  "3": {
    "time": "23:00",
    "team": "Team 5 - Team 6"
  }
}

---

* The filter only handles primitives AFAICT; if your data includes datetimes for example you will need to stringify them first.

Answer 2

I found a Gist and forked it to make trivial changes for python 3.x compatibility and it worked perfectly for me:

https://gist.github.com/jerrydboonstra/bedb5ccbe6200b5d794447ed38aeb342

It has an example.json file that acts as the template with "{{ variable }}" tokens, and a runner.py which composes the output from the template and an input data structure.