AMI Script
AMI Script Introduction
Overview
AMI has an embedded, versatile language which is a combination of the well-known C/Java style, SQL and String Templating languages. Here is a very quick example of the incredible versatility of the language which does a cross-database join:
1{
2 //Inspect the ACCOUNTS table of two pre-defined databases (qadb & proddb).
3 //Two temporary in-memory tables are created (qatemp and prtemp)
4 String table = "ACCOUNTS";
5 CREATE TABLE qatemp AS USE datasource=qadb EXECUTE SELECT * FROM ${table};
6 CREATE TABLE prtemp AS USE datasource=proddb EXECUTE SELECT * FROM ${table};
7
8 //Gather some statistics, Note that the results of the queries on the in-memory tables are
9 //stuffed into the local variables
10 int qaCnt = select count(*) from qatemp;
11 int prCnt = select count(*) from prtemp;
12 int bothCnt = select count(*) from qatemp, prtemp where qatemp.id == prtemp.id;
13
14 //Alert the user with some statistics, including the number of users in both qa and prod
15 session.alert("Found ${qaCnt + prCnt} ${table} rows: ${qaCnt} in qa and ${prCnt} in prod. \
16 There are ${bothCnt} ids existing in both");
17
18 //Create a Temporary ProdOnly table that contains all users in prod but not qa.
19 //This table can be used in the visualization layer
20 Create table ProdOnly as select * from prtemp where !(id in (select id from qatemp));
21}
In this introductory language you can see the interoperability of SQL and a procedural style language. The ${...} syntax is used for easily creating dynamic strings and SQL statements.
Key Features
AMI Script is a versatile language designed for optimized "on the fly" compilation, meaning expressions can be re-compiled and evaluated in an instant. Some key features of the language:
- Type Safe - All variables have well defined types, ex: String vs. Integer
- Compile Time Binding - Code is compiled before execution to minimize runtime issues
- Procedural - Custom methods (with overloading), along with a set of predefined procedures
- Object Oriented - Objects in AMI are represented as such in Ami Script
- Event Driven - Executes based on events, such as running a datamodel or clicking a button
- Implicit Casting - Casting from one type to another can be explicitly or implicitly defined
- Embedded SQL - For accessing external sources and tables directly in the language. See below
- String Templating - Ability to template code for reuse. See below
- Task Swapping - Tasks can be exported runtime for out-of-band execution. (separate thread)
Embedded SQL is an SQL-style language extension to AMI Script with near-full SQL support along with several enhancements geared towards data preparation for the visualization of data. All of these functions act on an in-memory database. It also supports "calling out" to external datasources. Some key features are:
- Query Clauses - For filtering, joining, and grouping data on the in-memory database
- Data Modification - For adding, updating, and deleting data on the in-memory database
- Schema Modification - For adding, editing, and deleting tables/columns on the in-memory database
- Use … Execute - For seamlessly running code on external datasources and storing results to the in-memory database
- Advanced Data Preparation and Analysis - For complex data calculations, not typically available on a standard SQL database
- Lambda Support -The ability to run AMI script code inside SQL. For example, to run a code snippet per row returned in a table
- Nested Queries - The in(…) clause supports full nested queries, and also supports multiple column in-clauses
- Intelligent Indexing - As joins, in-clauses, etc. are evaluated, indexes will automatically be spun up depending on data size and cardinality
AMI Script Template syntax is an extension of AMI Script with the aim of simplifying dynamic text generation, useful for creating dynamic HTML, dynamic SQL, etc.
- Embed Ami Script inside Text - Embed AMI script inside text allows for easy dynamic text generation
- Embed Ami Script inside SQL - Embed AMI script inside SQL allows for easy dynamic SQL generation
- Injection Protection - Properly escaping control characters, quotes, etc.
- Conditional Templating - Ability to do conditional & loop based template construction
AMI Script
Variable Naming
Case-sensitive combination of a-z, A-Z, _, 0-9. The first character must not be a number. Alternatively, variables can be wrapped in back ticks (`), in which case all characters are supported, allowing for accessibility with external language references. Note: `test` is equivalent to test. Variables are declared with the following syntax (the value will default to null if an expression is not supplied):
type variable_name [ = expression][, variable_name [ = expression] ... ];
Literals (Constants)
digits | Int constant |
0xdigits | Hex Int constant |
0digits | Octal Int constant |
digitsL | Long constant |
0xdigitsL | Hex long constant |
0digitsL | Octal long constant |
digits.digits | Float constant. Note, scientific (aka 'E') notation is supported |
digits.digitsD | Double constant. Note, scientific (aka 'E') notation is supported |
"chars" | String constant. Use backslash (\) to escape quotes, backslash (\\) and other control chars, ex: \n. A string can be defined over multiple lines using a trailing backslash on all but the last line, note that preceding white space is trimmed out for subsequent lines. Please see String Templating section for details on ${…} syntax |
\n\r\t<space> | White space (token delimiter) |
Reserved Type
All reserved types' values are immutable. Null is a valid value
Long | 64 bit signed whole number |
Double | 64 bit signed float |
String | Variable length string of UTF 8 characters |
Integer | 32 bit signed whole number |
Float | 32 bit signed float |
Byte | 8 bit signed whole number |
Boolean | boolean (true or false) |
Number | Base class for Long,Double,Integer,Float,Byte |
UTC | Timestamp with millisecond precision (since unix epoch) |
Examples
Integer i = 32;
Double d = 4.3, f = 132.d; //declaring multiple variables
Integer j = 0xaabcc;
String test = "hello world";
String Test = "Hello \
world"; //Multi line
Integer lng = 100000000000L;
Boolean flag = true;
Byte some_thing = null;
Casting
Variables are strongly typed, but casting can be explicit or implicit. If a cast fails, the result evaluates to null. Also, equality operations (==, <=, >=, >, <, !=) will also auto cast when possible as well.
Examples
String test1 = 123; //implicit cast from integer to string
String test2 = "0x123";//implicit cast from hexidecimal integer to string
Double val = "123.32"; //implicit case for string to double
Boolean b = "true"; //implicit cast from string to boolean
String val = (Integer) 123.d; //explicit cast from double to integer, then implicit cast to String
Boolean c = "test"; //Evaluates to null because the implicit cast from string to boolean is invalid
Byte d; //Evaluates to null by default
Variable Scoping
Scopes are defined using statement blocks, which are denoted by curly brackets: {…}. Within a statement block, variables are only visible after they are declared. When statement blocks are nested inside other statement blocks, all variables declared in the outer block are visible within the inner block as long as the variable is declared before the inner block. A variable cannot be re-declared when another variable with the same name is visible.
Examples
{
Byte a;
//a is visible;
Byte b;
//a & b are visible;
{
Byte c;
//a, b & c are visible;
Byte d;
//a, b, c & d are visible;
{
Byte e;
//a, b, c, d & e are visible;
}
//a, b, c & d are visible;
}
//a & b are visible;
Byte a; //Compiler error: a already exists and can not be redefined
}
Standard Operators (in order of significance)
math_expressions
(evaluates to num_var type)
num_expr | * | num_expr | multiplication |
num_expr | / | num_expr | division |
num_expr | % | num_expr | modulus (remainder) |
num_expr | + | num_expr | addition for numbers |
num_expr | - | num_expr | subtraction |
binary_expressions
(evaluates to num_int_expr type)
int_expr | ~ | int_expr | bitwise NOT |
int_expr | << | int_expr | signed bit shift left |
int_expr | >> | int_expr | signed bit shift right |
int_expr | >>> | int_expr | unsigned bit shift right |
int_expr | & | int_expr | bitwise AND |
int_expr | ^ | int_expr | bitwise XOR |
int_expr | ! | int_expr | bitwise OR |
boolean_expressions
(evaluates to boolean)
! | bool_expr | boolean (logical) NOT | |
expr | < | expr | less than (strings are case sensitive) |
expr | > | expr | greater than (strings are case sensitive) |
expr | <= | expr | less than or equal to (strings are case sensitive) |
expr | >= | expr | greater than or equal to (strings are case sensitive) |
expr | == | expr | equal to (strings are case sensitive) |
expr | != | expr | is not equal to (strings are case sensitive) |
expr | =~ | regex | matches regular expression (see Java Pattern for details) |
expr | !~ | regex | does not match regular expression (see Java Pattern for details) |
expr | ~~ | simple_pattern | simplified text matching. See Simplified Text Matching section. |
expr | instance of | type | does expr evaluate to supplied type (eg: would casting succeed) |
expr | && | expr | logical and. Short circuits if left expr returns false. Short circuits if right expr is const and false |
expr | expr | logical or. Short circuits if left expr returns true. Short circuits if right expr is const and true |
string_expressions
(evaluates to boolean)
string_expr+ expr | String concatenation. Null values equate to empty string ("") |
special_expressions
(expression) | Force Order of operations |
array_of_type[num_expr] | Array index referencing |
(type)statement | Explicit cast to type. Note that a failed cast returns null. Also, casting is automatic, so it's not necessary |
bool_expr ? tr_expr : fl_expr | If bool_expr returns true, then evaluates to tr_expr, otherwise evaluates to fl_expr. |
Examples
Long n = (10 + 5) * 2; //n is 30
n += 5; //n is 35
n = n/2; //n is 17
Boolean b = n>10; //b is true (n is still 17)
b =! b; //b is false
b = (b || true) && true;//b is true
String test = "hello" + "world"; //test is helloworld
Assignment & Declaration
special (evaluates to type)
new type(param[, param ...]) | Object Creation |
type var_name | Variable declaration (Value is defaulted to null) |
type var_name = expression | Variable declaration and assignment |
Variable declaration and assignment | Variable declaration and assignment |
assignment (evaluates to num_var type)
num_var | ++ | combined post-increment assignment | |
++ | num_var | combined pre-increment assignment | |
num_var | -- | combined post-decrement assignment | |
-- | num_var | combined pre-decrement assignment | |
num_var | *= | num_expr | combined multiplication assignment |
num_var | /= | num_expr | combined division assignment |
num_var | %= | num_expr | combined modulus assignment (remainder) |
num_var | += | num_expr | combined addition assignment |
num_var | -= | num_expr | combined subtraction assignment |
Flow Control
Complex AMI scripts can be achieved by combining multiple statements into statement blocks. A statement block is wrapped in curly brackets {...} and each statement is separated with a semi-colon;. Then, the execution of statement blocks can be determined using a variety of flow control statements including if, if/else, for, while, and for-each-loops. Note that for and while loops can use the continue and break directives as well. Examples
{
Integer count = 10;
Integer value = 1;
for(Integer i = 0; i < count; i++){
value *= 2;
}
if(value < 2000) {
session.alert(value); //alert shows 1024
} else
session.alert("High value"); // this is never hit. Also note: single statements don't need to be in {...} statement blocks
while(true){
value -= 10;
if(value < 1000)
break;
}
//value is now 994
}
statement block
{ statement [;statement ...] [;]}
** Evaluates to return value if a return_statment exists, otherwise it evaluates to the last statement's evaluation. If no statements, evaluates to Void
for loop
for(initialization;termination_expression;increment) statement
initialization: statement that is called first, used to declare variables with local scope
termination_expression: statement that must evaluate to boolean, false terminates
increment: statement called after each completion of the loop's statements
statement: called once per loop. Note:
- continue_statement will exit the statement and proceed to increment statement
- break_statement will exit the for loop entirely
**Evaluates to Void
for each loop
for(type_expr varname:iterable_or_array) statement
iterable_or_array - values to iterate over in order
statement: called once per loop. Note:
- continue_statement will exit the statement and proceed to increment statement
- break_statement will exit the for loop entirely
**Evaluates to Void
while loop
while(bool_expression) statement
same as for(;bool_expression;) statement
**Evaluates to Void
if statement
if(bool_expression) statement
if bool_expression evaluates to true then execute statement
**Evaluates to Void
if else statement
if(bool_expression) statement else statement2
if bool_expression evaluates to true then execute statement otherwise executestatement2
**Evaluates to Void
return statement
return statement
** Evaluates to statement's value. Will cause call stack to be popped out to caller function.
break statement
break
** See for_loop, for_each_loop and while_loop
continue statement
continue
** See for_loop, for_each_loop and while_loop
function call
function(param[, param ...])
**Evaluates to return_type value
member method call
target_object.function(param[, param ...])
**Evaluates to return_type value
Objects
There is a predefined set of object types (classes). These are used to facilitate the interaction between AMI Script and various backend components, user inputs, dashboards, etc. Additionally, data structures are represented as objects and can be instantiated using the new operator. Objects can only be accessed via methods (not member variables).
The Session Object
session - each user that is logged in has a session object (which is of type Session). This is the "starting point" for access to the user's information, along with various dashboard components.
Common Data Structures
List: A zero-indexed Array List implementation
Set: A set of unique values
Map: A Linked Hash Map implementation
Table: A 2-dimensional table with the same characteristics as a SQL style table. Tables have a title. Columns are named and strongly typed.
Tableset: A map of tables. Can be loosely thought of as a mini database.
Examnples
session.alert("this is a test");
session.log("something worth logging);
List l = new List("apple", "orange", "berry", "banana");
l.splice(1, 2, "strawberry");
String s = l; //s is now apple, strawberry, banana
Map m = new Map(1, "one", 2, "two", 3, "three"); //map contains three key/value pairs, ex: 2 = "two"
Commenting
Commenting is identical to the commenting syntax in Java & C++. There are two types of comments: single-line and multi-line.
comments (single-line)
//comments<linefeed>
comments (multi-line)
/*comments*/
Examples
/*multi-line
comment using the
slash stars*/
//single-line comment
AMI Script Constants
AMI Script can reference constants just like other variables. Note the constant's values can not be changed using AMI Script. The standard naming convention for global variables is 2 under bars (__) followed by an all uppercase name. Constants can be defined in three ways:
- At the server level (via a config file loaded at startup, typically local.properties)
Typically the syntax is: amiscript.variable.varname=value
(note that if value is a string, is must be surrounded in double quotes, see LITERALS for details)
For example, this could reside in your local.properties:
#My custom variables
amiscript.variable.__COMPANY_NAME="Acme Technologies"
amiscript.variable.__COPYRIGHT=2017
- At the user level (using the authentication plugin, or the access.txt file)
Typically the syntax is: amiscript.variable.varname=value
(note that if value is a string, is must be surrounded in double quotes, see LITERALS for details)
For example, this could reside in your access.txt file for user demo (with password demo123):
demo|demo123|amiscript.variable.__GROUP="infosec"|amiscript.variable.__SECURITY_LEVEL=12
- There are also pre-determined variables defined by the AMI Web system:
- __USERNAME: (String) the username of the logged in user
- __SESSIONID: (String) a unique id for the web session
- __TIMEZONE: (String) the timezone of the user's session
- __ADDRESS: (String) the ip address of the user's browser
- __LOADTIME: (Long) the time in milliseconds that the user started the session
- session: (Session) the session object representing the user's session
Hint: Through the front end in editor mode you can view all declared constants via: Menu Bar -> Dashboard -> Variable Table
SimplifiedText Matching
Overview
The simplified text matching finds text entries using a simple pattern. When included in AMI script, it takes the form: var ~~ "my pattern"
Rule 1, the general case: the pattern supplied will match any text that contains the pattern, case insensitive. For example, assuming word is a variable let's consider the expression: word ~~ "AB". This will return true if word is cab or ABC or tabs but will return false if word is apple.
Rule 2, special symbols: If you need to look for symbols besides letters, numbers or white space, prefix with a back slash. For example, to look for the word P&L, the expression is word ~~ "P\&L". Note this includes the following list of symbols: *^$.[]()~|&!?:
Rule 3, Additional syntax: There are many special symbols that can be used for more advanced searches, for example the star (*) means anything, so "f*g" will match any thing that has an f followed by g, such as 3FORGE. See the following sections for more advanced syntax.
Syntax The following syntaxes are used to do "basic" matching against expressions with some text.
1. Simple Matching Expression
Matches any text that contains the supplied pattern regardless of case. The following characters are reserved and may be embedded in the expression:
- * - Anything: Can be expanded to include any text.
- ^ - Starts With: Symbolizes the "start" of the text. In other words, if an expression starts with ^, then the text must start with the following expression. Please note that carets (^) not at the start of the expression will be treated as literals. (When used in conjunction with the tilde (~), the hat should follow the tilde.)
- $ - Ends With: Symbolizes the "end" of the text. In other words, if an expression ends with a dollar sign($) then the text must end with the preceding expression. Please note that dollar signs not at the end of the expression will be treated as literals.
- . - Any single char.: Will match to one character of any value.
- [charlist] Any of Chars.: Match any one of the characters enclosed within the brackets.
- ~: Case Sensitive. If the first character, indicates that matching characters should be case sensitive.
- Escaped chars: When referencing the following characters literally, they must be escaped with a back slash (\): *^$.[]()~|&!?:
- Literals: These are simply letter-for-letter matches to the expression
2. 'Exact Expression'
Matches the exact expression supplied in quotes (other than an escaped quote using \' ). This would be similar to using an equals ( == )
3. /regex/<options>/
This will use the java.util.regex package to match against regular expressions (see the java.util.regex.Pattern for details on building regular expressions). The <options> allow you to control behavior of how the regular expression operates. Simply supply the options you would like to apply in a continuous string between the 2nd and last forward slash (/). The following options are supported:
- q - Canonical Equivalence: When this flag is specified, two characters will be considered to match if and only if their full canonical decompositions match. The expression "a\u030A", for example, will match the string "?" when this flag is specified. By default, matching does not take canonical equivalence into account.
- f - Full Line: When this flag is specified, the text must fully match the regular expression. If this option is not specified, any portion of the text matching the expression will be considered a match.
- i - Case Insensitive: By default, case-insensitive matching assumes that only characters in the US-ASCII charset are being matched. Unicode-aware case-insensitive matching can be enabled by specifying the u (Unicode case) flag in conjunction with this flag.
- c - Comments: White space is ignored, and embedded comments starting with # are ignored until the end of a line.
- l - Literal: White space is ignored, and embedded comments starting with # are ignored until the end of a line.
- d - Disable Dot All: If the disable dot all flag is set then the dot expression (.) will match any character except a line terminator. By default, the dot expression only matches line terminators.
- m - Multiline: In multiline mode the expressions ^ and $ match just after or just before, respectively, a line terminator or the end of the input sequence. By default, these expressions only match at the beginning and the end of the entire input sequence.
- u - unicode Case: When this flag is specified, case-insensitive matching, enabled by the i (case insensitive) flag, is done in a manner consistent with the Unicode Standard. By default, case-insensitive matching assumes that only characters in the US-ASCII charset are being matched.
- x - Unix Lines: In this mode, only the '\n' line terminator is recognized in the behavior of ., ^, and $.
- v - Not Matching: Reverses the matching result. Expressions that normally would match will not match and expressions that would not normally match will match.
- n - No Error: If the regular expression is invalid do not throw an exception, instead just always return false, unless the v flag is also set, in which case always return true.
4. <null>
Null String - providing null will result in a matcher which matches nothing.
5. <Empty String>
Providing an empty (zero length) string to the Matcher method will result in a matcher which matches only the empty string.
Conjugated Syntaxes
The following syntaxes allow for more complicated pattern matching. Nested syntaxes may also be conjugated syntaxes. For nested syntaxes which are basic syntaxes open parenthesis [ ( ] and close parenthesis [ ) ] must be escaped with a backslash [ \ ].
6. (Expression): The parenthesis can be used to force order of operations, important when mixing or [ | ] and and [ & ] clauses.
7. Expression1|Expression2: If either of the nested expressions match this expression will return true.
8. Expression1&Expression2: If both of the nested expressions match this expression will return true.
9. !Expression: Inverts the result of the expression.
10. (Expression) ? (ExpressionTrue) : (ExpressionFalse): If the Expression matches, return result of ExpressionTrue Otherwise return result of ExpressionFalse.
Sample Expressions
Expression | Examples Match | Examples Mismatch | Comments |
---|---|---|---|
FORG | 3Forge | Frg | Partial text, Case insensitive match |
~FORG | FORGE | forge | Partial text, Case sensitive match |
^FORG$ | FORG | FORGE | Exact text, Case insensitive match |
~^Forge$ | Forge | forge | Exact text, Case sensitive match |
!FORG | 3FOR | 3FORGE | Escape partial text, case insensitive |
!^FORG$ | FORGE | FORG | Escape exact text, case insensitive |
[RB]ob | Bobby | lob | Using char list |
[A-Z]ob | mob | 7ob | Char list range |
3F*ge | 3Forge llc | Forge llc | Wild chars |
3f.ge | 3Frge | 3Forge llc | Match single char |
Acme llc\. | Acme llc. | Acme llcc | Escaping special chars |
robert | ted | Or clause partial match | |
rob | robert | Or clause, exact match | |
bob&rob | robob | bobob | And clause partial |
'Exact' | Exact | exact | Exact case sensitive text match |
a | Empty string, only matches empty string | ||
a?bb:s |
Abba test |
acca
tent |
If contains 'a', must contain 'bb'
If doesn't contain 'a', must contain 's' |
af | ab | Must contain a or b. Also must contain e or f | |
def | DEF | Must be exactly 'abc' or 'def' | |
/reg/ | regular | Regular | By default, regex must be case sensitive |
/reg/i/ | Regular | whatever | I option makes case insensitive |
3F
reg |
3Forge
Regular |
Mixing regex and non regex |