CS622 Assignment 2 - Sean Rawson

Requirements

Java 11
Jackson 2.13
JUnit 5

Running the program

The main entry point for the program is in class Main. The easiest way to try it out is to load the project in your IDE and run Main.

Program Design

This program is designed to implement two main areas of functionality. Those areas are file merging and file searching. JSON files are used for sample data, so the output of any search actually contains a list of objects of type JsonNode. This is one of the main return types used by Jackson when reading JSON strings or trees.

File Merging

File merging is accomplished by reading, line by line, through each file in a list of files. As each line is read, it is written to the destination file. This functionality is implement somewhat naively, with no permissions checking or validation of file formats. Invalid formats will be caught by the JSON processing code.

File Searching

With JSON files used for the project's sample data, the Jackson library was used to read in json strings and/or trees. The main class providing search functionality is JsonSearcher.

The searcher will attempt to process a JSON file as a list of JSON strings, one per line. If that method fails, it will attempt to process the file as a tree, meaning the format of 'pretty printed' JSON.

Every JsonSearcher has several private member variables that can be initialized via the constructor or set with the corresponding accessor methods.

JsonSearcher.startingField specifies the top-level field of each JsonNode that should be considered the parent of the desired search area. The thinking is that often JSON objects will be returned by an API with some metadata attached to them. For example:

{
    "source": "APIv2",
    "url": "https://api-v2.my-site.com"
    "data": {
        "key1": "value1",
        "key2": "value2"
    }
}

If we're only interested in nodes that are children of the "data" node, we can specify startingField in the JsonSearcher constructor. If the specified starting field doesn't exist in an object, then the entire object will be searched instead.

JsonSearcher.matchMethod specifies either KEYWORD or PATTERN matching for searches. This should be used carefully, as it can produce some unexpected results. If PATTERN is used and the search pattern is specified as a keyword (i.e. "value" instead of ".*value.*"), only nodes with exact matches will be returned. Strings like "value1" or "here is a value" will not match.

JsonSearcher.ignoreCase allows for case sensitive or case insensitive matching. Case sensitive matching is enabled by default.

Future Improvements

There are several areas that could be improved in this program.

File validation is essentially performed by the search functions, which fail when invalid formats are encountered. The code could be refactored to move that validation to another class. The search code would be simplified somewhat if there was an assumption of properly formatted files.

The search capabilities are not very fine-grained currently. They could be expanded to provide for searching only specific fields and/or searching some fields by pattern and some fields by keyword at the same time. One example might be searching for all projects that match a keyword and have a close date within a certain date range.

At present, only the "funds_raised_percent" and "close_date" fields are printed from each matching record. Allowing the user to select the output fields and customize the output format would be valuable additions.