Parsers

The parsers object provides various parsers for common data formats.

JSON5

The parsers.json5 function parses the JSON5 format. JSON5 is an extension to the popular JSON file format that aims to be easier to write and maintain by hand (e.g. for config files).

In general, parsing a JSON file as JSON5 does not cause harm, but it might be more forgiving to syntactic errors. In addition to JSON5, JSON repair is applied if the initial parse fails.

• JSON5 example

{
    // comments
    unquoted: "and you can quote me on that",
    singleQuotes: 'I can use "double quotes" here',
    lineBreaks: "Look, Mom! \
No \\n's!",
    hexadecimal: 0xdecaf,
    leadingDecimalPoint: 0.8675309,
    andTrailing: 8675309,
    positiveSign: +1,
    trailingComma: "in objects",
    andIn: ["arrays"],
    backwardsCompatible: "with JSON",
}

To parse, use parsers.JSON5. It supports both a text content or a file as input.

const res = parsers.JSON5("...")

YAML

The parsers.YAML function parses the YAML format. YAML is more friendly to the LLM tokenizer than JSON and is commonly used in configuration files.

fields:
    number: 1
    boolean: true
    string: foo
array:
    - 1
    - 2

To parse, use parsers.YAML. It supports both a text content or a file as input.

const res = parsers.YAML("...")

TOML

The parsers.TOML function parses the TOML format. TOML is more friendly to the LLM tokenizer than JSON and is commonly used in configuration files.

# This is a TOML document
title = "TOML Example"
[object]
string = "foo"
number = 1

To parse, use parsers.TOML. It supports both a text content or a file as input.

const res = parsers.TOML("...")

JSONL

JSONL is a format that stores JSON objects in a line-by-line format. Each line is a valid JSON(5) object (we use the JSON5 parser to be more error resilient).

data.jsonl

{"name": "Alice"}
{"name": "Bob"}

You can use parsers.JSONL to parse the JSONL files into an array of object (any[]).

const res = parsers.JSONL(file)

XML

The parsers.XML function parses for the XML format.

const res = parsers.XML('<xml attr="1"><child /></xml>')

Attribute names are prepended with ”@_“.

{
    "xml": {
        "@_attr": "1",
        "child": {}
    }
}

front matter

Front matter is a metadata section at the head of a file, typically formatted as YAML.

---
title: "Hello, World!"
---

...

You can use the parsers.frontmatter or MD to parse out the metadata into an object

const meta = parsers.frontmatter(file)

CSV

The parsers.CSV function parses for the CSV format. If successful, the function returns an array of object where each object represents a row in the CSV file.

const res = parsers.CSV("...")

The parsers will auto-detect the header names if present; otherwise you should pass an array of header names in the options.

const res = parsers.CSV("...", { delimiter: "\t", headers: ["name", "age"] })

PDF

The parsers.PDF function reads a PDF file and attempts to cleanly convert it into a text format. Read the /genaiscript/reference/scripts/pdf for more information.

DOCX

The parsers.DOCX function reads a .docx file as raw text.

INI

The parsers.INI parses .ini files, typically used for configuration files. This format is similar to the key=value format.

KEY=VALUE

XLSX

The parsers.XLSX function reads a .xlsx file and returns an array of objects where each object represents a row in the spreadsheet. The first row is used as headers. The function uses the xlsx library.

const sheets = await parsers.XLSX("...filename.xlsx")
const { rows } = sheets[0]

By default, it reads the first sheet and the first row as headers. You can pass a worksheet name and/or a range to process as options.

const res = await parsers.XLSX("...filename.xlsx", {
    sheet: "Sheet2",
    range: "A1:C10",
})

VTT, SRT

The parsers.transcription parses VTT or SRT transcription files into a sequence of segments.

const segments = await parsers.transcription("WEBVTT...")

Unzip

Unpacks the contents of a zip file and returns an array of files.

const files = await parsers.unzip(env.files[0])

HTML to Text

The parsers.HTMLToText converts HTML to plain text using html-to-text.

const text = parsers.HTMLToText(html)

Prompty

Prompty is a markdown-based prompt template format. GenAIScript provides a parser for prompty templates, with a few additional frontmatter fields to define tests and samples.

basic.prompty

---
name: Basic Prompt
description: A basic prompt that uses the chat API to answer questions
---
system:
You are an AI assistant who helps people find information. Answer all questions to the best of your ability.
As the assistant, you answer questions briefly, succinctly.
user:
{{question}}

To parse this file, use the parsers.prompty function.

const doc = await parsers.prompty(file)

Code (JavaScript, Python, C, C++, Java, …)

The parsers.code function parses source code using the Tree Sitter library. It returns an AST (Abstract Syntax Tree) that can be used to analyze the code.

// the whole tree
const { captures } = await parsers.code(file)
// with a query
const { captures } = await parsers.code(file, "(interface_declaration) @i")

The tags query is a built-in alias for the tree-sitter tags query that is made available in most tree-sitter libraries.

const { captures } = await parsers.code(file, 'tags')
```

## Math expression

The `parsers.math` function uses [mathjs](https://mathjs.org/) to parse a math expression.

```js
const res = await parsers.math("1 + 1")

.env

The parsers.dotEnv parses .env files, typically using for configuration files. This format is similar to the key=value format.

KEY=VALUE

fences

Parse output of LLM similar to output of genaiscript def() function. Expect text to look something like this:

Foo bar:
```js
var x = 1
...
```

Baz qux:

Also supported. …

Returns a list of parsed code sections.

const fences = parsers.fences("...")

annotations

Parses error, warning annotations in various formats into a list of objects.

• GitHub Actions
• Azure DevOps Pipeline

const annotations = parsers.annotations("...")

tokens

The parsers.tokens estimates the number of tokens in a string for the current model. This is useful for estimating the number of prompts that can be generated from a string.

const count = parsers.tokens("...")

validateJSON

The parsers.validateJSON function validates a JSON string against a schema.

const validation = parsers.validateJSON(schema, json)

mustache

Runs the mustache template engine in the string and arguments.

const rendered = parsers.mustache("Today is {{date}}.", { date: new Date() })

jinja

Runs the jinja template (using @huggingface/jinja).

const rendered = parsers.jinja("Today is {{date}}.", { date: new Date() })

tidyData

A set of data manipulation options that is internally used with defData.

const d = parsers.tidyData(rows, { sliceSample: 100, sort: "name" })

GROQ

Apply a GROQ query to a JSON object.

const d = parsers.GROQ(
    `*[completed == true && userId == 2]{
  title
}`,
    data
)

hash

Utility to hash an object, array into a string that is appropriate for hashing purposes.

const h = parsers.hash({ obj, other }, { length: 12 })

By default, uses sha-1, but sha-256 can also be used. The hash packing logic may change between versions of genaiscript.

unthink

Some models return their internal reasonings inside <think> tags.

<think>This is my reasoning...</think>
Yes

The unthink function removes the <think> tags.

const text = parsers.unthink(res.text)

Command line

Use the parse command from the CLI to try out various parsers.

# convert any known data format to JSON
genaiscript parse data mydata.csv