The POM project consists of a set of utility classes and methods. Many classes are self explanatory and some have the
documentation provided at a class level. The users of this module may also refer to the test cases in src/test folder.
Below is a high level overview of the key classes present in this module.
Package com.aexp.acq.unify.base.utils
BaseUtils class
This class consists of static helper methods. Please refer to the Javadoc documentation for these methods for further details.
BlockOnOfferQueue class
This class is used as the queue supplied to an Executor Service so the Executor Services blocks when an item is submitted for execution and there is no space in the queue to accept the item. In Java, there is no way to block in such a scenario and hence this implementation. In this implementation, the offer method which is used by the Executor Service is overridden and implemented using a put method which blocks. If we do not do this, then the alternatives are either an exception is thrown or a rejected item handler needs to be defined and the task executed on the current thread. The first is not acceptable and the second will lead to a multi threading situation where we specify a single thread in the pool which is not what we want.
ErrorMap class
A class used to store the error strings stored as key value pairs.
ERRORS_BASE class
A collection of error messages for classes defined in this module.
ErrorTuple class
An class representing an error in terms of code, message details and is retryable.
UnifyException class
The base class for exceptions hierarchy.
Package com.aexp.acq.unify.utils.stats
This package contains classes that can be used to instrument an application. It provides a mechanism to collect the information for the maximun, minimum and avarage response time for a function over one or more time durations.
StatsReceiver
This is the class that receives stat events. Stat is a short form for performance statistics. A StatsReceiver object
may contains one or more StatRecorders.
StatsRecorder
This class represents a stats recording abstraction. A recorder is an entity that is provided stat events to record.
This is an abstract base class which is extended for specific statistic recordings. Each StatsRecorder contains
multiple TimeSlice objects. Each time slice object specifies information of the time slice and stores the summary of
stats for that time slice.
ServiceStatsRecorder
This class extends StatsRecorder and is responsible for recording the performance statistics of services across
specified time slices. It is provided events of the type ServiceStatEvent to process.
StatEvent
This is an abstract base class representing a stat event.
ServiceStatEvent
A concrete class representing a stat event for measuring the performance of a service. Each object stores the name of the service, the timestamp at which the event was created, the response time measured and whether it was a case of an error or not.
StatSummary
This is the abstract base class for representing summary information.
ServiceStatSummary
This class is used to store the summary information for a specific service over a time slice. This class stores the following information for a service for a time slice:
- Minimum response tome
- Maximum response time
- Average response time
- Number of calls made
- Number of error calls
- Total response time
Setting up stats recording for an application requires the following to be done:
Initialize StatsReceiver
statsReceiver=new StatsReceiver(int numThreads,int queueSize);int numThreads specifies the number of threads that will be used by StatsReceiver to process incoming events. Usually
a single thread should work.
int queueSize specifies the size of the queue to hold the events waiting to be processed. Usually a size of 8 to 16
should work
Register Stat Recorders
First create a new recorder object.
ServiceStatsRecorder recorder=new ServiceStatsRecorder(int timeSlices,long timeSliceDurationInSec);
statsReceiver.registerRecorder(recorder);int timeSlices specifies the number of time slices to measure in the past
long timeSliceDurationInSec specifies the duration of each time slice in seconds
For example, (3, 259200) will mean "create 3 time slices, each of 3 days duration". Using this, the recorder will at any point of time, maintain 3 time slices, one of which will be current
Create a Stats Event
ServiceStatEvent e=new ServiceStatEvent(name,now,time,isError);String name is the name of the service
long now is the time stamp
long time is the response time of the service
boolean isError specifies if an error occurred or not
Send the event to the stats receiver
statsReceiver.accept(new Task(e));Read the API documentation.
Go through the unit test cases in the source code. Unit test cases are available in the location src/test
Provide us feedback. We would love to hear from you. We can be reached here:
Deepak Arora, deepak.arora5@aexp.com, GitHub: @deepakarora3
Shamanth B Chandrashekar, shamanth.b.chandrashekar@aexp.com, GitHub: shamanth9
Deepika Sidana, deepika.sidana3@aexp.com, GitHub: deepikasidana89
Jyotsna Gandhi, jyotsna.gandhi1@aexp.com, GitHub: jodube21
Shailendra Bade, shailendra.bade@aexp.com, GitHub: finaspirant
Phil Lundrigan, phil.lundrigan@aexp.com, GitHub: lundriganp
A special thanks to Phil Lundrigan for his tremendous leadership, support and encouragement.
We welcome Your interest in the American Express Open Source Community on Github. Any Contributor to any Open Source Project managed by the American Express Open Source Community must accept and sign an Agreement indicating agreement to the terms below. Except for the rights granted in this Agreement to American Express and to recipients of software distributed by American Express, You reserve all right, title, and interest, if any, in and to Your Contributions. Please fill out the Agreement.
Any contributions made under this project will be governed by the Apache License 2.0.
This project adheres to the American Express Community Guidelines. By participating, you are expected to honor these guidelines.
Package com.aexp.acq.unify.base.decision_table
This package implements a decision table functionality using Excel.
The excel based decision tables can be completely used from within an IDE and can be configuration managed (GIT). The Excel files are stored as resource files just like any other source code artifact, checked in / out of the repository and version controlled.
The decision table provide the following functionality:
- Passing of input data as a map of key value pairs
- Ability to add columns and rows as required
- "First match" and "All Matches" policies. First match policy returns when the first match is found. All match policy evaluates to the last row and returns results corresponding to all matching rows
- "Return None" and "Return Default" no match policies. This policy is applicable in case of no match found. In the case of return none, no values are returned while in the case of return default, the contents of the default row are returned back
- Standard operators =, <=, >=, <> that work on Boolean, double, integer, long and string data types
- A "Like" operator for string data type using which a cell can be evaluated against multiple values
- Ability to return contents of multiple columns for the matched rows
- For a return cell, ability to specify a function to be executed whose return value will be returned to the caller. To use this feature, enter the value of a return cell as "# full class path" and define a class which implements the "InvokableCell" interface. For example, enter "# com.americanexpress.unify.utils.decision_table.TestInvokableReturn". If the row is matched, then this class will be instantiated and execute method called on this class. The execute method will receive the map of input values passed to the decision table as well as a custom object which is up to the user of the decision table to define
Below is an example of a decision table implemented in Excel:
A decision table is created as below:
DecisionTable dt=DecisionTable.fromExcel("/base/DTTest1.xlsx");The input values are collected in a map (Map<String, String>) as one or more key value pairs and passed into the decision table for evaluation as below.
Map<String, String> values=new HashMap<>();
values.put("score","100");
values.put("yob","1974");
values.put("code","4GG");The response of a decision table is a list of key values pairs in the form List<Map<String, RetDTCell> as below:
List<Map<String, RetDTCell>>list=dt.evaluate(values);
for(Map<String, RetDTCell> map:list){
map.forEach((colName,dtCell)->System.out.println("Col name -> "+colName+", value -> "+dtCell.getValue()));
System.out.println();
}The list can be traversed to get the individual rows returned from the decision table using standard Java for loops.