Lark

Grammar

The grammar defines all the rules of your language and how to generate a parse tree.

Lark accepts grammar in extended baccus naur form or EBNF.

rule_name:
	regex patterns to match or "string"

Example (for a json parser):

() → grouping | → or [] → optional (zero or more occurrences) * → also means zero or more occurrences + → one or more instances

value: dict | list | STRING | NUMBER | "true" | "false" | "null"

list : "[" [value ("," value)*] "]"

dict : "{" [pair ("," pair)*] "}"

pair : STRING ":" value

We can also import rules, using %import. In the above example, we can import STRING and NUMBER from lark’s common library.

// arrow (->) is used to rename a rule, ESCAPED_STRNG is renamed as STRING
%import common.ESCAPED_STRING -> STRING
%import common.SIGNED_NUMBER -> NUMBER

We also need to deal with white spaces and new-lines. In our case we can simply ignore it using %ignore

// import whitespaces
%import common.WS
 
// ignore it
%ignore WS

One more thing to note is that will generating syntax tree, lark will automatically ignore any unnamed string literal. Named literal are ignore if prefixed with underscore _.

Lark will not ignore any named or unnamed regex even unless prefixed with underscore _.

e.g., "[" [value ("," value)*] "]" → here the brackets and commas are ignored in the syntax tree as they are unnamed string literals.

This would mean, lark would ignore "true" and "false" etc. We can fix that be doing this:

value: dict
	| list
	| STRING
	| NUMBER 
	| "true" -> true
	| "false" -> false
	| "null" -> null

We can use ? to reduce a node if it has only one child. For example, if a node has only one child, it makes sense to just remove that node.

Node value -> Node child → Node child

In our example, the value will have only one child. That means the value node is redundant.

?value: ...

---- data
{"key": ["item0", "item1", 3.14]}
 
---- old syntax tree
value
  dict
    pair
      "key"
      value
        list
          value	"item0"
          value	"item1"
          value	3.14
          
---- reduced syntax tree
dict
  pair
    string	"key"
    list
      string	"item0"
      string	"item1"
      number	3.14
      true

Generating the syntax tree

## pip install lark
from lark import Lark
 
parser = Lark("contents of the .lark file")
text = '{ "name": "samyabrata maji", "socials": [{ "name": "github", "username": "samyabrata-maji" }, {"name": "twitter", "username": "sammaji15"}] }'
 
st = parser.parse(text)
print(st.pretty())