A general-purpose single-header C logging library and parser for event-based logs. (Incomplete)
{#} | Time | Hash | Subject | Author | # | (+) | (-) | GPG? |
---|---|---|---|---|---|---|---|---|
46 | 26 Dec 2018 17:01 | 9046b13 | Create LICENSE | Joshua | 1 | 674 | 0 | N |
Object | Latest Commit Subject | Time | Hash |
---|---|---|---|
src | Fix RecursiveMkdir, so that it works cross-platform. | 26 Dec 2018 23:21 | 921db36 |
static | Logo | 25 Dec 2018 19:58 | 5a96206 |
.gitignore | Update .gitignore | 25 Dec 2018 23:56 | bb9606b |
LICENSE | Create LICENSE | 26 Dec 2018 17:01 | 9046b13 |
README.md | Update README.md | 26 Dec 2018 13:25 | ece9cde |
index.js | Update index.js | 26 Dec 2018 13:25 | 26d3797 |
package-lock.json | Create package-lock.json | 25 Dec 2018 21:19 | 478c0cd |
package.json | Update package.json | 26 Dec 2018 13:25 | 97a345f |
Advanced node.js logging for advanced programs.
Most loggers available only use a linear method of logging; there are verbosity levels and tags to narrow searches, but everything is still on the same level or plane nonetheless. The package lognestmonster
is a similar type of logger, but it allows you to create multiple layers (or "nests") of log statements. This is useful because, although you may have to put in some extra organizational work on the code side, the log results are much more clean and thorough, boosting workflow efficiency. The user has absolute control over the way log statements are pushed to their log files.
Why should I use this? There are many projects that create insane amounts of data, almost impossible to sift through without a helper program. The purpose of this logging system is to be a time-saver. Although it takes more time to put it in place, it's almost immediately made up with the performance gain through using the new log format. One thing that's unique about this logger compared to others is that it allows multiple queues to be made, in turn allowing you to split up your data. For example, a Node web server could keep one log file that records everything that the backend does while it uses another to record user or traffic information for analytics. Parsing software could be used to read either one.
There are 4 classes offered by the package: Logger
, the driving device that organizes everything; Queue
, the class that actually pushes the log statements to their respective file; Statement
, the actual log data (timestamp, verbosity, tag/invoker, message); and Event
, a nest layer for Statement
s or other Event
s.
The following subsections assume that lognestmonster
has been require
d by node.js with the following code:
const Logger = require("lognestmonster");
Creates and organizes your queues. Logger
is the exported package. Logger.Logger
is the Logger class.
Logger.Logger()
var MyLogger = new Logger.Logger(Object config);
Where Object config
defaults to:
{"name": "Logger", "locations": {"node": "./log/node"}}
name
serves no purpose other than identification. locations
is used to create new queues, taking each key as the queue name and each value as the queue output location.
Logger.Logger.queue()
MyLogger.queue(string name);
This returns the appropriate Queue
object for the provided name
, assuming it was created with the Logger
object.
Manages log Statement
s and Event
s and how they're written to the final log file. Note that these are implicitly created with Logger.Logger
when the locations
object is properly provided in the config
parameter of the Logger.Logger
constructor.
Logger.Queue()
let MyQueue = new Logger.Queue(string name, string location);
// note that parameters are the same format as key-value
// pairs in `config.locations` of `Logger.Logger(config)`
This creates the queue, taking name
to be used as its ID and location
as the path to where the log file should be created.
Logger.Queue.push()
MyQueue.push(Statement statement);
MyQueue.push(Event event);
MyQueue.push(string verbosity, string tag, string message); // implicitly creates a `Statement` object
This adds log items or nests to the to-write queue. This returns the Queue object.
Logger.Queue.write()
MyQueue.write();
This appends every queue value to the log file, emptying the queue. This returns the Queue object.
This is the base log item where written log data is actually held.
Logger.Statement()
let MyStatement = Logger.Statement(string verbosity, string tag, string message);
The timestamp value is created automatically by the constructor.
This is the proper name for a nest. Essentially, it's just an array that can hold other Event
objects and Statement
objects, creating a tree.
Logger.Event()
let MyEvent = Logger.Event();
let MyEvent = Logger.Event(Event event);
let MyEvent = Logger.Event(Statement statement);
let MyEvent = Logger.Event(string verbosity, string tag, string message);
Any arguments given are passed to this.push()
.
Logger.Event.push()
MyEvent.push(Event event);
MyEvent.push(Statement statement);
MyEvent.push(string verbosity, string tag, string message);
This message pushes an Event
or Statement
as items in the nest. In the case that 3 strings are given as arguments, a Statement
is implicitly created. This returns the Event object.
When creating a Statement
, you can pass anything you'd like for the first verbosity
string, although there are some ones preset by the package:
Logger.INFO
Logger.DEBUG
Logger.VERBOSE
Logger.VERYVERBOSE
Logger.WARNING
Logger.ERROR
These verbosity levels can be used to narrow down your search results when parsing the log files.
Here's an example of how somebody would initiate the Logger, create and push items to the Queue, and write them to the log file. PLEASE SEE THE NOTES ABOUT THIS EXAMPLE DOWN BELOW.
// Require the package
const Logger = require("lognestmonster");
// Creates the logger object. Placing the new Logger inside the package allows cross-file usage, so you only have to initiate once.
Logger.Overseer = new Logger.Logger({
name: "Overseer",
locations: {
"node": "./log/node" // Creates a queue named `node` that uses the path `./log/node`
},
});
// Pushes a statement directly to the `node` queue
Logger.Overseer.queue("node").push(Logger.INFO, "PROCESS", "Process started");
// Creates a new event and pushes a Statement to it
let LoadEvent = new Logger.Event();
LoadEvent.push(Logger.INFO, "INIT", "Acquiring needed top-level packages.");
// Creates a new event and pushes multiple Statements to it
let LowerNestedEvent = new Logger.Event();
LowerNestedEvent.push(Logger.INFO, "INIT", "Loading fs...");
LowerNestedEvent.push(Logger.DEBUG, "INIT", "fs loaded.");
LowerNestedEvent.push(Logger.INFO, "INIT", "Loading http...");
LowerNestedEvent.push(Logger.DEBUG, "INIT", "http loaded.");
LowerNestedEvent.push(Logger.INFO, "INIT", "Loading jsonwebtoken...");
LowerNestedEvent.push(Logger.DEBUG, "INIT", "jsonwebtoken loaded.");
LowerNestedEvent.push(Logger.INFO, "INIT", "Loading lognestmonster...");
LowerNestedEvent.push(Logger.DEBUG, "INIT", "lognestmonster loaded.");
// Pushes the Event LowerNestedEvent to the Event LoadEvent
LoadEvent.push(LowerNestedEvent);
// Pushes another statement to LoadEvent
LoadEvent.push(Logger.INFO, "INIT", "Finished.");
// Pushes LoadEvent (a nest that consists of [Statement, Event, Statement] now) to the write queue
Logger.Overseer.queue("node").push(LoadEvent);
// Queue should now look like this: [Statement, Event]
// Writes the queue, effectively emptying it into the log file
Logger.Overseer.queue("node").write();
The above code creates the following JSON-like log output in the designated file:
{"timestamp":"2018-12-26T18:08:37.654Z","verbosity":"INFO","tag":"PROCESS","message":"Process started"}
[{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Acquiring needed top-level packages."},[{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading fs..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"fs loaded."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading http..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"http loaded."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading jsonwebtoken..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"jsonwebtoken loaded."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading lognestmonster..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"lognestmonster loaded."}],{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Finished."}]
To see how to parse this data (putting it into proper JSON), see the Log Format section.
Large projects
If your project spans multiple files, you could easily place your Logger object into the package itself, allowing the same Logger object to be accessed by other parts of your project. This is seen in the sample code with Logger.Overseer = new Logger.Logger(...)
.
Repetition
There is some repetitive code; specifically, Logger.Overseer.queue("node")
. Do note that this actually results in a Queue
object, so you could easily make this its own variable like this:
let NodeQueue = Logger.Overseer.queue("node");
NodeQueue.push(...).write();
Queue pushing and writing
When you're pushing multiple objects to a queue, be wary that they will stay there until the queue is written. Because everything is its own class, what you're really pushing is a reference to the real object, so you can make changes to a pushed object after it has been pushed. As such, it is entirely possible to prematurely write a queue before an already-pushed Event
or Statement
is finished. It's good practice to immediately write the queue immediately after it has been pushed. Perhaps in the future there will be some functionality where a Statement
or Event
can be directly written instead of placed in the queue.
Push order
The developer has control over everything here. The order of your log file is the order that you push to your Event
s and Queue
. If I were to push a new Statement to LoadEvent
in the middle of the pushes to LowerNestedEvent
, it would show up on the log first because LowerNestedEvent
isn't itself pushed to LoadEvent
until later in the code. The order of the log file is 100% logical and in the developer's control, so it would be wise to write your code neatly with this in mind. This logging package doesn't do anything you don't tell it to.
Logs are controlled by Queue
objects. Placed in the folder they're told, the logs should follow the ISO datetime format and have a .log
file extension. Here's the name of a sample log file: 2018-12-26T18-08-37-653Z.log
Regarding format, logs follow a JSON-like format, as follows:
{"timestamp":"2018-12-26T18:08:37.654Z","verbosity":"INFO","tag":"PROCESS","message":"Process started"}
[{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Acquiring needed top-level packages."},[{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading fs..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"fs loaded."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading http..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"http loaded."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading jsonwebtoken..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"jsonwebtoken loaded."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Loading lognestmonster..."},{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"DEBUG","tag":"INIT","message":"lognestmonster loaded."}],{"timestamp":"2018-12-26T18:08:37.655Z","verbosity":"INFO","tag":"INIT","message":"Finished."}]
Because of the way the Queue objects push it (the most efficient way regarding computing power with changing/appending to files), you will have to do a bit of tweaking to get it in proper JSON format:
This could be easily automated. Once finished, you get proper JSON (below), where each object {}
is a Statement and each wrapping table []
is an Event (excluding the outermost one).
[
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "PROCESS",
"message": "Process started"
},
[
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "INIT",
"message": "Acquiring needed top-level packages."
},
[
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "INIT",
"message": "Loading fs..."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "DEBUG",
"tag": "INIT",
"message": "fs loaded."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "INIT",
"message": "Loading http..."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "DEBUG",
"tag": "INIT",
"message": "http loaded."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "INIT",
"message": "Loading jsonwebtoken..."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "DEBUG",
"tag": "INIT",
"message": "jsonwebtoken loaded."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "INIT",
"message": "Loading lognestmonster..."
},
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "DEBUG",
"tag": "INIT",
"message": "lognestmonster loaded."
}
],
{
"timestamp": "2018-12-26T05:55:23.360Z",
"verbosity": "INFO",
"tag": "INIT",
"message": "Finished."
}
]
]
The above can be taken in by parsing software and create something similar to the following:
[[LOG FILE NAME]]
[[LOG DATE]]
[[LOG FILE SIZE]]
[[LOG ITEM COUNT]]
TIMESTAMP - INFO - PROCESS - Process started
v 3 ITEMS
TIMESTAMP - INFO - INIT - Acquiring needed top-level packages.
v 6 ITEMS
TIMESTAMP - INFO - INIT - Loading fs...
TIMESTAMP - DEBUG - INIT - fs loaded.
TIMESTAMP - INFO - INIT - Loading http...
TIMESTAMP - DEBUG - INIT - http loaded.
TIMESTAMP - INFO - INIT - Loading jsonwebtoken...
TIMESTAMP - DEBUG - INIT - jsonwebtoken loaded.
TIMESTAMP - INFO - INIT - Finished.
This is exciting! You still have the ability to narrow down your results with timestamps, verbosity levels, and tags/invokers, but now you have the organization with collapsable nests to not be overwhelmed with large amounts of data at the same time.
Developer - Joshua 'joshuas3' Stockin <joshstockin@gmail.com> (https://www.github.com/joshuas3)
Name - Patrik 'Patrola' Xop (https://github.com/PatrikXop)