In this exercise, you are asked to create a webpage that allows you to search for news
information using the Google News API and the results will be displayed in card format. The
Google News API is documented here:
2.1. Description of the Home Page
A user first opens a page, called index.html (or any valid web page name), where they can read
top headlines provided by the Google News API along with specific headlines from CNN and
Fox News. The top headlines are displayed in a sliding format (Figure 1). The top headlines from
CNN (Figure 2) and Fox News (Figure 3) are displayed in a card format. A word cloud for
frequently used words is displayed as shown in Figure 1.
Figure 1: Initial Home Screen
Figure 2: CNN layout on home page
Figure 3: Fox News layout on home page
To access the Google News API using Python, you should use the Python Client Library
You need to use Flask Python framework to make all the API calls. Using xmlhttprequest or
Define routing endpoints and make your API call from the python backend. The recommended
tutorial for Flask and more importantly, routing, can be found at the following link:
To initialize the Google News API client in Python, use this function call:
newsapi = NewsApiClient(api_key=[API_KEY])
The function call to request the news top headlines can be coded in Python as in the following
’, category=’business’, language=’en’, country=‘us’)
The URL that will be generated by the Python Client Library looks as follows:
The process to create your API key is explained in section 3.1. When constructing the python
request required to obtain the homepage news, you need to provide three parameters:
• The first parameter is your API_KEY that you create in section 3.1 for initializing the client
• The second parameter is the news source you want the headlines from. For the generic case, there
is no source to be used. For CNN and Fox News, the source parameter values should be ‘cnn’ and
• The maximum number of articles to be returned by the API per page are to be restricted using the
‘page_size’ attribute in the client request above and restricting it to 30 articles per page
The response of this Python request is a JSON-formatted object. Figure 4 shows a sample
response from the request. You need to parse this JSON object and extract some fields as
required. The data for the slides and the card layouts is from the corresponding JSON structure
as shown in Figure 5.
Figure 4: Sample JSON response from Google News API’s top-headlines endpoint
Figure 5: Sample JSON response from Google News API for slide and card layout
Slide/Card Data Data from Google News API response JSON
Image The value of key “urlToImage”
Title The value of key “title”
Description The value of key “description”
Table 1: Mapping JSON data and card display
All news articles are clickable and open the actual news article on a new tab using the URL
returned in the “url” key of the JSON structure above.
NOTE: The headlines in the slide are limited to 5 headline articles but NOT necessarily the
first 5 headline results, but the first 5 that have all keys (author, description, title, url,
urlToImage, publishedAt and source with its inner keys) present. The same applies for the
card format for CNN and Fox News (4 each).
For the word cloud, you need to find the top 30 frequent words from the title of the news article
sorted from most frequent to least frequent words. This list of words should NOT contain
stopwords. You can download the list of stopwords from the following link:
Please download the “stopwords_en.txt” file for the homework. To display the list of frequent
words, the d3-cloud library is to be used. The following links should suffice in helping you
display the word cloud:
The tabs at the left of the home screen – Google News and Search – have a hover effect as
shown in Figure 6 and in the video.
Figure 6: Hovering on Search while current selection is Google News
Clicking the search button displays a search form as shown in Figure 7.
Figure 7: Search Form
2.2. Description of the Search Form
The search form (Figure 7) consists of the following fields:
• Keyword: A required field that cannot be empty. It is an edit box the allows to enter
keywords for searching news about the keyword. The default value is empty.
• From: A required field for the date from which the articles are to be fetched. The default
value when the form is initially displayed is the date 1 week ago from the current date.
• To: A required field for the date to which the articles are to be fetched. The default value
is the current date.
• Category: A drop down to select among categories that are given to you in section 3.3.
Default selection is all.
• Source: A dynamically changing drop down that changes depending on the selected
category. Default selection is all.
• Search button: To allow search functionality only when a keyword is present. The input
field validation is done on clicking the search button and appropriate errors need to be
displayed. Missing fields should generate alerts as shown in Figure 8. An example of
valid input is shown in Figure 9.
• Clear button: Resets the form to default values and clears search results.
Figure 8: Error message for missing keyword
Figure 9: Valid Input example
Another error condition is if the user input time is not in the mm/dd/yyyy date format or the
from date is after the to date as like From – 02/11/2020 and To – 02/07/2020. The error
message is shown in Figure 10.
Figure 10: Alert to show incorrect time
display is mm/dd/yyyy, it is up to you to convert it to any format you like for handling it in your
Figure 11 and Figure 12 show the error conditions when the default values of a date are
removed completely and are missing some part of the date, respectively.
Figure 11: Date is missing completely
Figure 12: Date is missing the day and year component
The dropdown menu for sources is updated dynamically using the sources endpoint of the
Google News API.
sources = newsapi.get_sources()
Anytime a category is selected in the search form, a call is made to fetch all sources available for
that category. The URL that will be generated looks like this:
The parameters you need to use for the above request are:
• API_KEY: The API_KEY generated for the Google News API.
• category: The category selected from the dropdown in the search form.
• language: The language parameter will always have the value ‘en’.
• country: The country parameter will always have the value ‘us’.
2.3. Displaying Results
In this section, we explain how to use the form data to construct a Python request to the Google
News API and display the result in your page.
construct a URL and send it to your Python endpoint where the request to query the Google
News API’s “everything” endpoint is made. The URL has the following format
This is an example call using the Python client library:
domains=’bbc.co.uk,techcrunch.com’, from_param=’2017-12-01′, to=’2017-12-12′,
language=’en’, sort_by=’relevancy’, page=2)
The function call will require the following parameters:
• API_KEY: Your API_KEY to initialize the API client.
• q: The query keyword retrieved from the search form.
• from_param: The date From retrieved from the search form as start date.
• to: The date To which the articles are to be fetched. Retrieved from the search form.
• sources: The selected source from the drop down in your search form.
• language: The language parameter will ALWAYS have the value of ‘en’.
• page_size: Restrict to 30 articles per page retrieved.
• sort_by: Sort the results by the time they were “publishedAt”.
The data returned from the API call is to be displayed as shown in Figure 14.
The JSON data returned has the same mapping as shown in Table 1 on page 4. Only the articles
that have author, title, description, url, urlToImage, publishedAt, source and inner key of source
(name) keys present with a non-empty value are to be displayed, if any attribute is missing or
empty or blank, do not display it to a user.
If more than 5 articles returned in the results, you should display the first 5 articles along with
the Show More button underneath. When clicking the button, a maximum of 10 more articles
should be displayed and the Show More button changes to Show Less as shown in Figure 15.
After displaying the Show More button, at the end of all results (upto 15) there will be a Show
Less button clicking which will again display just 5 results.
NOTE: The Google News API has a restriction on how far back in time it allows the start date
for a search to be. As a result of this, you are expected to handle this error case by displaying the
error message returned by the API, as shown in Figure 13.
Figure 13: Google News API returns this error
Figure 14: Search Results example
Figure 15: Show More changes to Show Less
The card descriptions need to be truncated with an ellipsis (“triple dots”) without cutting out any
word in the middle so as to only span a SINGLE line. As can be seen in the reference video and
in Figure 16, on hovering over the card, the styling changes as shown.
Figure 16: Card styling difference on hover
Each card in the display is clickable and results in additional information being displayed as
shown in Figure 17.
Figure 17: Expanding the result card
The mapping between the information populated in the result and the JSON object is shown in
Slide/Card Data Data from Google News API JSON response
Image The value of key “urlToImage”
Title The value of key “title”
Description The value of key “description”
Author The value of key “author”
Source The value of key “name” inside the key “source”
Date The value of key “publishedAt” converted to the
format of mm/dd/yyyy
Link to Original Post The value of key “url”
Table 2: Mapping between result card and JSON object
The “See Original Post” hyperlink opens the original post in a new tab. The close button (the
“X”) at the top right corner of the card returns the card to the collapsed state as shown below.
Figure 18: After pressing the X button in Figure 16
2.4. Saving Previous Inputs
In addition to displaying the results, your page should maintain the provided values to display the
current result. For example, if one searches for “Keyword: Book, From: 02/03/2020, To:
02/10/2020”, one should see what was provided in the search form and the corresponding results.
It follows that you need to retain the whole search box/input fields and buttons even while
displaying results/errors. This “saving” happens only when the window is open. You do NOT
need a database/Local Storage or any persistence mechanism.
3.1. How to get Google News API Key
Go to https://newsapi.org/docs.
Click on “get API key”. You should fill out the form as shown in Figure 19. The API Key
(Google News API key) will be displayed. Copy this key as you will use it in all the Google
News web service API calls.
Figure 19: Google News Register for API Key Form
3.2 Deploy Python file to the cloud (GAE/AWS/Azure)
You should use the domain name of the GAE/AWS/Azure service you created in HW #5 to
make the request. For example, if your GAE/AWS/Azure server domain is called
example.appspot.com or example.elasticbeanstalk.com or example.azurewebsites.net, the
following links will be generated:
GAE – http://example.appspot.com/index.html
AWS – http://example.elasticbeanstalk.com/index.html
Azure – http://example.azurewebsites.net/index.html
example in the above URLs will be replaced by your choice of subdomain from the cloud service
3.3 List of Category for dropdown
You should these categories in the respective dropdown in the Search form:
3.4 Useful Links
A. For Google News API
B. For Word Cloud
C. For Stopwords
D. For Flask/Python Backend
E Styling hints
• Styling a horizontal rule – https://www.w3schools.com/howto/howto_css_style_hr.asp
• Button styling – https://www.w3schools.com/css/css3_buttons.asp
• Shadow on a card:
4. Files to Submit
In your course homework page, you should update the Homework 6 link to refer to your new
initial web search page for this exercise (for example, index.html). Your files must be hosted on
GAE, AWS or Azure cloud service. Graders will verify that this link is indeed pointing to one of
the cloud services.
Also, submit your source code file to the GitHub Classroom repository so that it can be graded
and compared to all other students’ source code via the MOSS code comparison tool.
• All discussions and explanations in Piazza related to this homework are part of the
homework description and grading guidelines. So please review all Piazza threads, before
finishing the assignment. If there is a conflict between Piazza and this description and/or
the grading guidelines, Piazza always rules.
• You should not use JQuery or Bootstrap for Homework 6.
in a 4-point penalty.