Crate squeeze 

Squeeze

A library for extracting rich information from any text.

This crate provides finders for extracting structured data from text:

• uri::URI - Extract URIs/URLs/URNs as defined by RFC 3986
• cidr::Cidr - Extract CIDR notation (IPv4/IPv6 network ranges)
• codetag::Codetag - Extract codetags (TODO, FIXME, etc.) as defined by PEP 350
• color::Color - Extract colors (hex, rgb, hsl)
• datetime::Datetime - Extract ISO 8601 datetimes
• domain::Domain - Extract DNS-style domain names
• email::Email - Extract email addresses
• emoji::Emoji - Extract emojis and emoji sequences
• env::Env - Extract environment variable references
• handle::Handle - Extract social @handle mentions
• hash::Hash - Extract hashes (MD5, SHA-1, SHA-256, SHA-512)
• ip::Ip - Extract IP addresses (IPv4, IPv6)
• json::Json - Extract JSON objects and arrays
• jwt::Jwt - Extract JSON Web Tokens
• mac::Mac - Extract MAC addresses
• modeline::Modeline - Extract vim/vi/ex modelines
• path::Path - Extract file paths (absolute, relative, and home-relative)
• phone::Phone - Extract phone numbers
• semver::Semver - Extract semantic versions
• uuid::Uuid - Extract UUIDs
• mirror::Mirror - A passthrough finder that returns the entire input

The scanner::Scanner runs any set of finders over a line, or over a whole buffer of lines with scanner::Scanner::scan_buffer, in one pass: a prescan disables finders whose bytes are absent, SIMD classification and per-finder context gates keep finder calls to the positions where a match can start, and declarative run rules gate several finders at once from a single measurement of the digit and hex runs. See docs/performance.md in the repository for the architecture, the contracts finders follow, and how to measure.

Example

use squeeze::{uri::URI, Finder};

let finder = URI::default();
let text = "Visit https://example.com for more info";

if let Some(range) = finder.find(text) {
    println!("Found: {}", &text[range]);
}

Modules

cidr

codetag

Codetag finder implementation based on PEP 350.

color

datetime

domain

Domain name finder.

email

emoji

env

handle

Social handle finder.

hash

ip

json

Strict JSON finder.

jwt

mac

mirror

Mirror finder - a passthrough finder for debugging.

modeline

Vim modeline finder.

path

phone

scanner

semver

uri

URI finder implementation based on RFC 3986.

uuid

Structs

Anchor

Bytes that every match of a dispatch finder contains, and the bytes that may lie between a match’s start and its first anchor byte, see Finder::anchor.

AnchorCheck

When the anchor byte at pos is one of anchors, the byte at one of the offsets after pos (or at every offset, with all) must be one of bytes (and exist) for pos to be the first anchor byte of a match: an IPv4 address’s first . has another two to four bytes on, a UUID’s first - others 5, 10 and 15 bytes on, a MAC address’s first : others 3, 6, 9 and 12 bytes on.

ByteSet

A set of bytes, used by RunRule.

Memo

Per-line, per-finder scratch handed to Finder::try_at_memo/Finder::try_trigger_at_memo.

Run

A maximal run of one RunClass starting at a candidate position.

RunRule

A declarative rule on the run starting at a candidate position, see Finder::run_rules.

Runs

The digit, hex and word runs at a candidate position, measured once by the scanner and shared by every finder’s RunRules. The word run is only measured when a candidate finder has a RunClass::Word rule.

Then

A run required right after the byte that ends the previous run of a RunRule: class bytes numbering min..=max, followed by a byte in after (or the end of the input when after_end is set).

Enums

RunClass

The byte class a RunRule measures.

Constants

MAX_CHECKS

Most checks an Anchor can carry (one per group of anchor bytes).

MAX_CHECK_OFFSETS

Most offsets an AnchorCheck can list.

MAX_THEN

Most follow-up runs a rule can chain, see RunRule::then.

RUN_CAP

Longest run length the scanner distinguishes; longer runs are reported as this value with Run::capped set.

Traits

Finder

A trait for finding patterns in text.