Raw data entry is only the start of your digital quality management. Based on this data, you want to make decisions and steer your operation. These decisions rely on predefined business rules and logic. Imagine how effective and powerful your quality system would be if it could apply these rules and logic automatically. Well, that is exactly what AlisQI is capable of with the expression engine.

The calculations that you can add to analysis sets are powered by our expression engine. This versatile engine enables you to implement rich logic in your analysis sets. Combined with the workflows, this enables you to create business logic within AlisQI.

This article presents an overview of how to create and use advanced expressions in AlisQI. Due to the vast possibilities, this guide is necessarily incomprehensive. Please refer to the resources on the bottom of this page for full documentation of the underlying syntax and available features.

# Introduction

Let's start with a short introduction video on the power of the AlisQI expression engine.

# Syntax and features

The expression engine is based on Twig, an open-source templating engine.

The expression engine supports a rich syntax with many features, including functions, filters, tests, and operators.

Since the expressions run in a sandboxed environment there is no security risk. Users can never get to the core application code by means of the expressions engine.

### Variables

All result values are available in the expression engine as variables. These variables are defined in braces, just like the field name in the example below:
```{fieldname_}
```

During the evaluation of the expression, the variable contains the value of that particular field. Calculated values are available as well. This enables you to create complex, nested expressions.

### Operators

Twig provides a variety of operators for math, logic, comparisons, and other means.

Here are two examples to show you a few possibilities of operators.

1. if the user is known, display the username, display 'unknown user' otherwise

```'Hello, ' ~ ({user_} ? {user_} : 'unknown user')
```

2. if the remainder of a divded by b is non-zero, return b, else 1

```{a_} // ({b_} != 0 ? {b_} : 1)
```

### Filters

Variables can be modified by filters. Filters are separated from the variable by a pipe symbol ( `|`). Multiple filters can be chained. The output of one filter is applied to the next.

The 'upper' filter would convert 'diana' into 'DIANA'

```{name_}|upper
```

The 'capitalize' filter would convert 'diana' into 'Diana'

```{name_}|capitalize
```
See Twig documentation on filters

### Functions

Functions can be called to generate content. Functions are called by their name followed by parentheses () and may have arguments.
For instance, the date function can convert date(time)s. The following function call returns the current date minus 2 days:

```date('-2days')
```

### Tests

Tests can be used to test a variable against a common expression.

```{numericfield_} is odd
'pass' in {status_}
```

### Greater than and less then operators

Values with a `<` or `>` sign are inaccurate by nature. Calculating these values can be risky, as the outcome can be inaccurate. AlisQI aims to keep the data accurate. This can lead to not performing a calculation that cannot be resolved to an accurate value.

# Beyond result values

The expression engine offers more data than just the result values. You also have specifications, field visibility, attachments, and more available to use. All this to empower you to build rich business logic into your quality system.
The table below defines the entire context available in the expression engine.
 `{field_}` The value for a particular field in your analysis set. Can be `null`. For selection fields, this is the string value of the chosen option (e.g., "Pass"). `{fieldname_}.id` For selection fields only The option's static id, as integer numeric value. Use this to prevent changes to the option value (e.g., from "Pass" to "OK") from impacting the calculation. `{fieldname_}.option` For selection fields only This syntax allows you to retrieve all field values from a selection list option. See example #7 below. (Technically speaking, the `.option` is a hash of the selected option's values.) `{field_}.spec` The current specification values for this field. For numeric fields (including calculations), this is a hash with keys `min, innermin, goal, innermax, max`. Access any boundary using `{field_}.spec.min`. For selection fields, this is a list (array) of in-spec options (e.g., `['White', 'Transparent']`). `{field_}.offspec` and `{field_}.offspec_internal` For selection and numeric fields only Boolean value which indicates whether the value is outside of (internal) specification. `{fieldname_}.label` and `{fieldname_}.unit` The field's label and unit (if any, can be `null`). `{fieldname_}.decimals` and `{fieldname_}.default` For numeric fields (including calculations) only The field's number of decimals and default (if any, can be `null`). `{fieldname_}.visibility` The current field visibility state for this field. `{fieldname_}.operator` For numeric fields (including calculations) only The operator (< or >) set for this field (if any, can be `null`). `attachments` The list (array) of filenames, e.g., `['CoA.pdf', 'Delivery by pallet.jpeg']`.

# Examples

To demonstrate the power of the expression engine, we now go through some real-life examples. These are just meant to inspire you. The possibilities are endless. Dive into the resources and go nuts!
Expiry date
Calculate the expiry date by adding 3 months to the production date
```DATE_ADD_MONTH({productiondate_}, 3)
```
Alternatively, we could use the date filter to transform the production date string to a unix timestamp, add the necessary number of seconds, and convert back to a date string.
```({productiondate_}|date('U') + 3 * 30 * 24 * 60 * 60)|date('Y-m-d')
```

Logical expressions

If (condition) then (A) else (B)

Can be defined as:

```condition ? A : B
```
Use goal spec if no value was inputted
Use the inputted value, or fall back to the goal spec if no value was inputted. (Uses the ternary operator)
```{ph_} ?: {ph_}.spec.goal
```

Test if an attachment was uploaded

```attachments is empty
? "No attachments"
: ("Number of attachments: " ~ attachments|length)
```

List all pdf attachments

```attachments|filter(att => att ends with '.pdf')|join("\n")
```

Find the lower of two dates
Convert date1 and date2 to timestamps, determine the minimum value, are transformed back to yyyy-mm-dd.

```MIN({date1_}|date('U'), {date2_}|date('U'))|date('Y-m-d')
```

Actually, thanks to the formatting of date strings (yyyy-mm-dd), we can compare the date strings directly.

```MIN({date1_}, {date2_})
```
Suppose the field `assignee_` is a selection field based on the user list. The following will evaluate the chosen user's address.
```{assignee_}.option.email
```

Test if a result has off-spec values
You might want to trigger or deviation process in case a result has an off-spec value. The following expression tells you whether the result has an off-spec value:

```offspec|filter(s => s) is not empty
? "Has off-spec values"
: "No off-spec values"
```

# Custom calculations from Excel to LIMS

One of our customers, a medium-sized food factory, had a hard time replacing their Excel QC sheets with a standard solution. Not because they did not have a budget or time, but because of the complexity of their Excel sheets. They had many, many calculations in their sheets. The AlisQI expression engine provided the solution.

# Errors and warnings

The enormous flexibility offered by the expression engine does come with a price. It's possible to create calculations that will run into problems.

AlisQI will try to validate the calculation by evaluating it using dummy values, but it cannot possibly catch all errors. Therefore, a list of errors and warnings is displayed while creating the calculation, as well as in the result entry screen. Use the preview in the analysis-set management to test the calculation with different input without needing to actually store results.

### Examples

Consider the following example for a numeric calculation: `{\$batch_}`, where Batch is a text field. The syntax of this calculation is fine, but whether a batch number can be interpreted as a number depends on its value. "1234AB" will be interpreted as the number 1234, but "AB1234" cannot be interpreted as a number. AlisQI cannot predict what values will be used, so it can only warn that using a text field in a numeric calculation can cause issues.

Here's a more devious example, using a numeric field `a : 1 + ({a} ?: {})`. As long as the value for `a` is truthy (not empty and not 0), this calculation will evaluate to `1 + a`, which will always be a number. However, if the value for `a` is empty, the engine will try to add a number and a hash, which isn't defined and causes an error.

# Caveats and technical details

If you read Twig's documentation, you may have noticed a couple of discrepancies with the syntax of calculations. To understand the differences, let's dive into how calculations are evaluated.

First of all, calculation is enclosed in output delimiters: `{{ calculation }}`. Therefore, you cannot use Twig tags.

Next, the field variables (e.g., `{field_}`) are actually invalid syntax. AlisQI transforms these to `values['field_']`. Likewise, `{field_}.spec.min` is transformed to `spec['field_'].min`, etc. The sandbox context actually contains (only) the following variables: `attachments, values, id, operator, decimals, default, label, unit`. You can use these if you wish.

# Useful resources

Here are some suggestions for more information on the technology behind our expression engine.