S13-Words-for-the-Future-notebooks/working-with-datasets/json-making-datasets.ipynb

Making (mini) datasets using JSON files¶

Ways of thickening, layering, ..., text.

---

JSON?¶

JSON is an open standard file format that saves information in key, value paired objects.

It looks like this:

{
    "name": "XPUB1",
    "date": "26-10-2020",
    "number_of_students": 10
}

In a JSON file, you can store both strings and numbers, but also lists or another dictionary object.

Look for example how the number 10 is written without "s around it.

dictionaries & JSON files¶

Python uses a dictionary to store data in the same key, value paired way.

dict = {}

We will use a dictionary to store the data, and then save it as a JSON file.

In this way, we can add one (of multiple) layer(s) of value to a word ..., such as:

{
    "common": "English",
    "language": "communication system",
    "formal": 6,
    "semantic": ['language', 'formal', 'informat'],
    "semantic": { 
            "type": "word",
            "number of letters": 7
        }
}

In [1]:

# Adding a new key to the dictionary, assigning a string as value:
dataset['new'] = 'NEW WORD'
# or assigning a number as value:
dataset['new'] = 10 

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-1-f8b69ae2bb04> in <module>
      1 # Adding a new key to the dictionary, assigning a string as value:
----> 2 dataset['new'] = 'NEW WORD'
      3 # or assigning a number as value:
      4 dataset['new'] = 10

NameError: name 'dataset' is not defined

In [ ]:

# Printing the tag of a word in the dictionary
print(dataset['are'])

In [ ]:

# Printing all the keys in the dataset
print(dataset.keys())

Making a sample dataset¶

In [3]:

# This is sample data, a list of words and POS tags:
dataset = [('Common', 'JJ'), ('languages', 'NNS'), ('like', 'IN'), ('English', 'NNP'), ('are', 'VBP'), ('both', 'DT'), ('formal', 'JJ'), ('and', 'CC'), ('semantic', 'JJ'), (';', ':'), ('although', 'IN'), ('their', 'PRP$'), ('scope', 'NN'), ('extends', 'VBZ'), ('beyond', 'IN'), ('the', 'DT'), ('formal', 'JJ'), (',', ','), ('anything', 'NN'), ('that', 'WDT'), ('can', 'MD'), ('be', 'VB'), ('expressed', 'VBN'), ('in', 'IN'), ('a', 'DT'), ('computer', 'NN'), ('control', 'NN'), ('language', 'NN'), ('can', 'MD'), ('also', 'RB'), ('be', 'VB'), ('expressed', 'VBN'), ('in', 'IN'), ('common', 'JJ'), ('language', 'NN'), ('.', '.')]

In [4]:

# Making a dataset with only verbs
dataset = {}

for word, tag in data:
    if 'VB' in tag:
        dataset[word] = tag

print(dataset)

{'are': 'VBP', 'extends': 'VBZ', 'be': 'VB', 'expressed': 'VBN'}

Saving as JSON file¶

In [5]:

import json

In [6]:

out = json.dumps(dataset, indent=4)
print(out)

{
    "are": "VBP",
    "extends": "VBZ",
    "be": "VB",
    "expressed": "VBN"
}

In [7]:

f = open('json-dataset.json', 'w')
f.write(out)
f.close()

In [ ]:

In [ ]: