Tests have been written for classes in the types directory, cfe-01 sub-directory still has a few classes that require them

Description

Title
Use case or motivation behind the feature request

Optional wrappers are basically just null values, so its better to move away from them and use exceptions

Additional context

yegor

The "Parser" name doesn't conform into OOP naming practices, hence it should be name something like "Hub".
The parseDirs method inside the Parser class should be elevated into its own class and called PromiseList.
CFE-01 and RELP Parsers should be renamed to ...Promise (This will be a single class with a generic type parameter.

Description

So instead of extending the RelpPromise, there should be a field for that purpose

Description

There are 39 warnings about error-prone code, mostly NonApiType ArrayLists and missing locale declarations.

• Fix ArrayLists
• Fix StringCaseLocaleUsage
• Fix inconsistent capitalization
• Fix ZonedDateTime.now() default timezone
• Fix Default charset
• Fix HidingField in CFE01GenerateFile (extends RelpGenerateFile)
• Find out what to do with MHC exclusion in HubDirList

Description

The NullSafeJsonObjectBuilder doesn't deal with the fact null fields shouldn't be sent to CFE_18, so it should be removed, and the Exception thrown from Jakarta JSON JsonObjectBuilder should be caught.

ID <20>
Reporter <@kortemik>

java package name should start with:
groupId+artifactId
in the case of cfe-31 it should be
fi.failsafe.lms.cfe31
i.e. package httpController; would become package fi.failsafe.lms.cfe31.httpController;
please use IDE refactor rename for the package name and it will take care of the rest.

1. Logging

Currently sending requests only returns an HttpResponse object, which is then individually logged by the sender.

The plan is to log the response at the client level, and not use the same duplicate code after returning the HttpResponse object.

Logging is simple and universal with this:

if (response == null) {
    Typewriter.technical_error("Response was null");
} else if (response.statusCode() == 201 || response.statusCode() == 200) {
    Typewriter.response(response.body());
} else {
    Typewriter.errorResponse(response.body());
}

2. Generating requests and sending them

Currently:
Client is first initialized, then a request is generated from the client object and stored in a var, then client is called with that var.

Plan is to make that var into the client object, and just call void methods to send it, and add store the response inside the client object.

Description

Log a faultyReport if CFE-18 returns an error for a duplicate Host when sent
-> makes finding bad data easier from the source

Description

CFE-18 has new endpoints for adding Storages, so those need to be fixed:

1. Add Storage
2. Add Flow to Storage with the new Storage ID
3. Link Storage with Capture

Description

Currently GenerateFiles read HostConfigFile and parse, which shouldn't be their responsibility, so a new class is required.

Lines like these are going away as well:

if (!configuration.hostConfigFile.path.toFile().exists()) {
    throw new NullPointerException("HostConfigFile is null");
}

• HostConfig should be responsible for validating itself.
• GenerateFiles should ask HostConfig to validate itself before asking for correct flows and protocols

Description

As stated above, create a new ctor to remove code from the Client class:

return new JsonResponse(
    jsonResponse.id,
    jsonResponse.message,
    response.statusCode()
);

New version would be:

final JsonResponse jsonResponse = apiResponse.toJsonResponse();

Description

Currently the JsonResponse class is constructed twice in the Client.java file, once by marshaling it with Gson, and the creating a new JsonResponse with the message, id and the response.statusCode() method.

This will be refactored to use another class, ApiResponse, and using a toJsonResponse() method to convert the HttpResponse inside it to the JsonResponse class

Description
Currently the Client is created each time a request is sent, consuming resources and possibly disregarding the benefits of connection pooling.

TODO:

1. Refactor the Client so that it's declared in Main.java, and passed to proper classes
2. Get the endpoint URI from the CLI arguments and pass it to the Client instead of hard-coded one
3. [ ] Refactor error response handling so that it accepts string instead of the preformed JSON format

Description

Those classes currently don't have their own send methods, but are sent from Host class in a loop

Description

Going to switch from using Gson to the Jakarta JSON API (previously known as javax.json API)

Description

There are still some getters present, so they should be renamed to something more OO.

Regex search get.*\(.*\).*\{ results in 52 matches in 28 files.

Description

Add readme.adoc with sections mentioned in repo-template readme.

ID <17>
Reporter <@MoonBow-1>

There are some directories that are parsed, even though they are not CFE-01 integrations.
Some filters should be added to remove them from parsing.

TODO:

• #2

HostName field is not populated at the moment, but should be done

ID <38>
Reporter <@MoonBow-1>

Modifications to CFE-18 require a Hub to be sent before Hosts

The Configuration class should be divided into smaller classes, so they can be passed to other methods/classes more easily and efficiently.

Would be looking at a maximum of 8 classes.

https://stackoverflow.com/questions/12669497/using-enum-as-key-for-map/12669569#12669569

They should be replaced with primitive int

Description

TargetEnum shouldn't create a StorageFlow inside the send method and call the StorageFlow's send method after it.

Description

As stated, when comparing the two classes, there should be a result class

This was related to an old ticket for CFE-18, which has now been fixed, and it breaks the send method for CaptureMeta.

If the name has been marked as a duplicate, the response from the server will be the original name of the duplicate if the data is the same. This also applies to CaptureMetas with a different name, but the same data

Description

Sub-classes to move:

• HubDir
• Send

Description

This came up in a review, that JavaDoc is missing from most methods, so that would be great to add

Description

That method is currently found in RelpGenerateFile and overwritten in CFE01GenerateFile, but is missing from the interface

ID <36>
Reporter <@MoonBow-1>

Different types of generate files might be missing certain fields, update exclusions to look for the missing fields

Closely tied to #15

Methods should throw Exceptions instead of returning null values.

Title

Now that the #3 is pretty much done, RELP groups are in line with the amount of groups sent, but CFE-01 groups differ by 26.

• 17 of these are due to the missing Host configuration, and are listed in the missingHostDirList.txt file after running the program.
• The rest require more in-depth investigation for the reason

Setters should be removed and constructors should be used to make the class immutable where applicable

Description

Requested by @Nefarious46 to move to Log4J2 for better logging

TODO:

• Implement Log4J2
• Implement SLF4J on top of Log4J2
• Change formatting of logs to match company policy

Description

Related to #22

Just going to remove all usages of gson.fromJson() method.

ID <30>
Reporter <@MoonBow-1>

CFE-18 has been updated, and the endpoint which have been changed and now there is more of them.

Endpoints to do:

• storage/flow
• storage/capture
• host/relp
• host/meta/ip
• host/meta/interface
• host/group
• capture/relp - Only 1 hub works, rest are missing Omrelp
• capture/groups/linkage
• capture/group

Description

Methods:

RelpGenerateFile getArchiveGenerateFile() throws IOException, JsonParseException;

RelpGenerateFile getMasterGenerateFile(CaptureMetaMap masterCaptureMetaMap)
    throws IOException, JsonParseException;

should return different classes, e.g. ArchiveGenerateFile and MasterGenerateFile.

Describe the bug

Currently HostMetas are sent even if they contain fields which are null

Expected behavior

HostMetas with null fields are not sent and are reported in a file

Software version

1.0.0

Preparing CFE-31 for release.

Issues to fix:

• Move files from resources to somewhere else
• Change some of the configuration.properties values to CLI arguments
• Remove unused dependancy
• Install CFE-31 in test env
• Confirm connection to CFE-18

If a field is final, it can be made into a public field, removing the getter for it.

private ArchiveDir archivingDir;

public ArchiveDir getArchivingDir() {
  return archivingDir;
}

Can just be made into:

public ArchiveDir archivingDir;

Description

CFE-31 should read JSON files "near" their usage place, instead of reading it once and passing the JsonObject around to the required class.
Use case or motivation behind the feature request

Related issues

#32 Issue came up from this, but requires more refactoring than a single class

Additional context

Example of correct implementation was pointed to: https://github.com/teragrep/snw_01/blob/main/src/main/java/com/teragrep/snw_01/tableapi/APIResponse.java#71

ID <4>
Reporter <@MoonBow-1>

There should be a diagram explaining the workflow of the application, all the files it accesses and how it uses those files.
They should also be kept up to date.

Description

As the title says

Description

It came up that the resource path should probably be a field for the class instead of being hard-coded in the send() method.

Description

CFE-31 is going to be published to GitHub someday, so all the somewhat confidential information should be changed to made-up data wherever possible, mostly tests

Tasks:

• Refactor tests' configuration instances (contains env specific information)
• Remove FIXME/TODO that references an env specific file with its FQ path
• Figure out what to do with the JSON configuration files
• Change publicly routable IP addresses to localhost
• Add copyright

Description

CFE-18 changed a few endpoint, so now some of the responses don't contain an id field, but a more specific name for the id

To fix:

• CG
• HG

Description

Validate methods should validate self, and not be dependent on outside parameters

The parser currently looks for exact matches between the master and archive dirs, and prod has the dirs without the cfe-01 prefix.

The options for this would be:

• Add prefixes
• Change the parser to check for a dir name without the prefix