A token is a substring of a password.

Extract all substrings from the input Text. A substring has a minimum character length of 3 for performance and to prevent false positives for matches such as sequences and repeats.

Examples:

>>> map _tokenChars (allTokens "abcdef")
["abc","abcd","abcde","abcdef","bcd","bcde","bcdef","cde","cdef","def"]

Translate the characters of a Text value.

Given a function that translates a character into one or more characters, return all possible translations.

Examples:

>>> translateMap l33t2Eng "p111
["piii","plii","pili","plli","piil","plil","pill","plll"]

The known patterns we are searching for.

Information about how a token matches a specific match pattern.

All possible matches after various transformations.

Final mapping of a token to its lowest score.

Map of partially applied estimates.

A function that will produce an estimate once we know the estimates for other tokens. This is necessary to score repeat matches since they require looking up the score for a different token.

Estimate all of the given matches.

Estimate a single match.

A password and estimated guesses represented as a graph.

A node in a guessing graph.

An edge is a guessing graph.

Given a password and a user word list, produce graph edges that connect the characters of the password.

Brute force edges. In other words, the edges required to ensure there's a path in the graph from the start node to the end node.

Generate a guessing graph from the given password and user word list. In the guessing graph the nodes are the characters in the password and the edges are the estimated guesses.

Collapse a graph down to a single score which represents the estimated number of guesses it would take to crack the password.

Calculate the shortest path through a guessing graph. In other words, the cheapest path for guessing a password.

Type alias for a frequency database.

Type to represent a ranking.

Look up the given value in all configured dictionaries, transforming each input with the given function. The lowest ranked score is return if it is found.

Track a translated l33t speak token.

Translate a token from l33t, counting l33t characters.

Convert l33t characters to their English character mappings.

A Pattern is two Unicode characters next to one another in a password.

Direction of movement for adjacent characters.

Movement between characters.

Keyboard layers.

Information about how two characters are related to one another.

An adjacency graph (usually representing a single keyboard).

Find a pattern if it exists. If all characters in the given Text form a pattern in the given Graph then a list of matches will be returned.

Scoring information for adjacent characters.

Calculate the score for two adjacent characters.

Information about a found pattern.

Allow other code to access the token used in a pattern.

Helper function to check if a token forms a keyboard pattern.

Estimate the number of guesses needed for a keyboard pattern to be cracked.

Type alias to represent the distance between characters.

If the delta between all of the characters in the given text are the same, that delta is returned.

Estimate a sequence.

Uses the scoring equation from the paper and not from the other implementations which don't even use the calculated delta. The only change from the paper is to compensated for a delta of 0, which isn't accounted for in the paper.

A date as a triple.

Components of a found date (year, month, day).

If the given text wholly contains a date, return it.

Extract the date components of a Date record.

Estimate the number of guesses for a date match.

Deviations from the zxcvbn paper:

1. The other implementations limit the year multiplier to 20 so we do the same here.
2. The other implementations multiply by 4 when date separators are used in the token. We do the same.

Internal mapping of repeating tokens.

Type alias for a count of repeating tokens.

Generate a repeat map from an existing token map.

Test to see if the given token is repeated.

If a repeat is found, the number of occurrences is returned along with the full token representing the repeating sequence.

In other words, if the token passed in is "word" and in the map we find that the original password contains "wordword", we return 2 to indicate 2 repeats and the token that represents the sequence "wordword".

A type to control which dictionaries, keyboard layouts, etc. will be used when estimating guesses.

Type alias for a frequency database.

Default configuration for US English.

Access all configured dictionaries.

Add a custom list of words for frequency lookup. The words should be ordered from most frequent to least frequent.