Markdown to JSON processing badly handles empty values

[SuperMayo]

There is a possible improvement in b696398 in the way we convert markdown to JSON.

Bug:
If the value is empty (for example - identifier: ), the equivalent json will be

{
...
"identifier": "",
...
}

which is both ugly and nonstandard for JSON.
The convention would rather be to have no name/value pair at all or having the null value

Solution:
My improved code based on the actual one:

observations = []
record = dict()
for line in txt.split("\n")[2:]:
    if line[:2] == "##":
        print(line)
        if len(record) != 0:
            observations.append(record)
        record = dict()
    else:
        if not line.strip()=="":
            print(f'process "{line}"')
            lhs, rhs = (line[2:].split(":"))
            lhs = lhs.strip()
            rhs = rhs.strip()
            if rhs != ""
                record[lhs] = rhs
            else:
                record[lhs] = None

The python None value will translate to null when using json.dumps()
One could also just drop the else statement and simply get no name/value pair for this line.

[albop]

Makes sense. Can you submit a PR ?