December 2020 Summaries

Load testing is a crucial aspect of software development that ensures an application performs well under normal conditions, simulating multiple users accessing it concurrently. It's essential to distinguish between load testing and stress testing, as the latter aims to discover where a system starts to break, whereas load testing seeks to understand how the system performs under normal conditions. Load testing is often performed sporadically or at significant software milestones, and various tools exist to aid in automated testing, including Apache JMeter, LoadNinja, LoadRunner, and Eggplant. These tools help teams plan, set up, analyze, and measure user experience, identifying bottlenecks and areas for improvement, ultimately saving organizations from costly performance errors.

The Google Vision API with Python allows developers to automate the management of online accounts by solving captchas, which are images that prevent scripts from interacting with websites. To use this API, developers need to install Python on their server, get an API key, subscribe to the Google Vision API, use the API with Python, and validate the results. The process involves connecting to the API using the `http.client` library, sending a POST request with the image URL, and parsing the response as JSON. The code provides examples of how to loop through a list of sample captcha images and compare the detected string with the actual string. While the accuracy of the Google Vision API may not be perfect, it can help developers get past an image captcha after 1-3 attempts, making it a useful tool for automating online account management.

The eBay Price Analytics API is a powerful tool for searching and retrieving product information from top eCommerce portals such as Amazon and eBay. With this API, developers can automate price comparison across multiple marketplaces and gain valuable insights into product pricing trends. The API offers a freemium subscription model with 100 API calls per month available for free under the BASIC plan. To get started, developers need to sign up for a RapidAPI developer account, navigate to the search page in the RapidAPI Console, subscribe to the Price Analytics API, and verify its operation by triggering a POST Create session request and then polling the results using a GET Poll session results endpoint. The API returns product listing data in JSON format, which can be extracted and visualized using tools like Plotly. By leveraging this API, developers can create interactive charts to compare prices across multiple marketplaces, enabling them to make informed decisions about product pricing and inventory management.

The COVID-19 data API is a Freemium API that offers a free tier with limitations for testing purposes. To connect to the API, an API key must be obtained and passed between software applications and the API as a unique identifier and secret authentication token. The API key can be hidden in environment variables or binary files to prevent exposure in scripts. Two methods are used to hide and retrieve the API key: environment variables and binary files. Environment variables can be set on Windows OS by navigating to Control Panel, System and Security, System, and Advanced System Settings, while Mac OS / Linux OS requires modifying a .bash_profile file. Binary files can be created using a Python script that writes the API key to a binary file. The API key can be retrieved from environment variables or binary files using Python scripts. The COVID-19 data API offers various endpoints for retrieving data, including confirmed cases, critical cases, fatalities, and recoveries. By following these steps, developers can access the API's endpoints with their API keys hidden.

The Call of Duty API with Python can be used to gather metrics from the game, including score, experience points, kill/death ratio, level, and prestige. The API provides a five-step process for using it: installing Python, obtaining an API key, subscribing to the API, using the API with Python code, and charting the results to look for correlations. The script used in this example pulls data from the first 450 pages of the leaderboard on the "battle" platform, retrieves nearly six thousand records, and charts several metrics against time played. The results show a strong correlation between time played and experience points, but weaker correlations with score and kill/death ratio. The levels generally increase with time played, while the prestige reaches a maximum level of 100. The data suggests that success in the game may not be directly correlated with hours played, as some players have high scores despite playing for thousands of hours, and others have low scores despite playing for shorter periods.