zackattak01 / humantimeparser Goto Github PK

Currently, the value of the FirstParsedTokenIndex and LastParsedTokenIndex properties indicates the index of the token based on the sections of the raw text. This behavior should be changed so that the indexes represent the index of the last char in the raw text.

The current implementation of the tokenizer and parser is quite pitiful and is only going to worsen. Rewrite it use a Token class instead of an enum and remove the squashing of multiple tokens into one from the tokenizer. The tokenizer should tokenize and the parser should parse.

Provides for the case where the "value" part of the input string would be parsed as part of the date/time.

Unknown formats should no longer be ignored and treated as they simply dont exist.

For example: "does things in 1s"
The current tokenizer completely ignores "does", "things", and "in". These should be tokenized as an unknown token type.

The current date parsing is limited to the capabilities of DateTime.TryParse. I hope to enable month and day parsing

This project may actually deserve some xml docs

This project is rather slow (no surprise, I wrote it of course).

Benchmark results