dotdict-parser 0.6.0

A unified path parser in C for Python

Universal JS like path parser for python

Specification

1. Overview

This specification defines how to reference properties or indices in a nested data structure using a combined notation:

1. Dot notation segments: a.b.c
2. Bracket notation segments: ["a"]["b"][0]
3. Mixed notation: a["b"].c[0]

It also imposes constraints to avoid ambiguous or invalid dot notation:

• No leading dots (e.g., .something is invalid).
• No trailing dots (e.g., something. is invalid).
• No repeated consecutive dots (e.g., a..b is invalid).

---

2. Dot Notation Rules

1. Valid Identifiers Only:
   • Each dot-delimited token (segment) must be a valid identifier in the target environment.
   • This typically means alphanumeric characters plus underscores (e.g., myKey, user2, some_value).
2. No Leading Dot:
   • A path cannot begin with a dot. For example, .a.b is invalid.
3. No Trailing Dot:
   • A path cannot end with a dot. For example, a.b. is invalid.
4. No Consecutive Dots:
   • A path cannot contain .. (two or more consecutive dots). For example, a..b is invalid.

Examples of valid pure dot paths:

• user.name.first
• accountDetails.balance
• config.version

Examples of invalid pure dot paths (under this specification):

• .leadingDot (leading dot)
• trailingDot. (trailing dot)
• double..dots (multiple consecutive dots)

---

3. Bracket Notation Rules

Bracket notation allows any valid string or number inside square brackets:

["some.key"] ["another space"] [42]

3.1 String Keys

• Must be enclosed in quotes inside the brackets: ["any string"] or ['any string'].
• Permits spaces, dots, dashes, or reserved characters in the key.

3.2 Numeric Indices

• To represent arrays or list indices, use bracket notation with a numeric literal: [0], [10], etc.
• No quotes needed for a pure integer.

3.3 Constraints on Brackets

• Bracket notation does not have the same dot-based restrictions (no worry about leading/trailing dots).
• However, mixing bracket notation into a dot path must still obey the dot-notation constraints for the dot-delimited segments.

---

4. Mixed Usage

When dot notation and bracket notation are combined, the path can jump between the two. Here are the rules:

1. Dot notation segments must follow the dot rules (valid identifiers, no leading/trailing dots, no consecutive dots).
2. Bracket segments can appear anywhere in the path, typically after a valid dot segment or another bracket segment.
3. A path cannot be just a dot—there must be a valid segment or bracket after/before each dot.

Examples:

• a.b[0].c
  • a and b are valid identifiers in dot notation, [0] is a bracket for an array index, then .c is back to dot notation.
• user["personal.info"].preferences["color.theme"]
  • Dot segments: user, preferences
  • Bracket segments: ["personal.info"], ["color.theme"]

---

5. Unified Path Grammar (Informal)

Here is a conceptual grammar that respects the stricter dot-notation constraints:

Path := DotSegment ( ('.' DotSegment) | BracketSegment )*

DotSegment := ValidIdentifier

BracketSegment := '[' ( StringLiteral | NumberLiteral ) ']'

ValidIdentifier := (Alpha | '') (AlphaNum | '' )* # No dots, no leading digit, no spaces, etc., based on your environment's definition.

StringLiteral := '"' '"' | "'" "'"

NumberLiteral := Digit+

• Path starts with a DotSegment (this disallows a leading dot).
• A . must always be followed by another DotSegment (disallowing empty segments or multiple consecutive dots).
• You can insert a BracketSegment at any point to handle special keys or numeric indices.

No leading/trailing dot is enforced by starting with DotSegment and only permitting subsequent '.' DotSegment pairs. No consecutive dots is enforced by requiring a segment name after each '.'.

---

6. Example Paths Under This Specification

1. Pure Dot Notation, All Valid

   • employee.profile.name
   • config.version.major

2. Mixed Dot & Bracket

   • a.b[0].c
   • root["user data"].info["age.range"][2]

3. All Bracket

   • ["root"]["user data"]["something.with.dots"][42]
   • Even though you can reference everything in brackets, if you switch to dot notation, any segment must be a valid identifier.

4. Invalid

   • .leadingDot.segment (leading dot)
   • segment. (trailing dot)
   • a..b (multiple consecutive dots)
   • root[""][""] is technically valid bracket usage for empty strings, but if you wanted empty string segments in dot notation, that’s disallowed here.

---

7. Traversal Semantics

Implementations using this path specification will:

1. Parse the path into a sequence of segments (each being either a valid dot identifier or a bracketed key/index).
2. Traverse the nested data structure in order:
   • For a DotSegment ( DotSegment ): move to the sub-key DotSegment.
   • For a BracketSegment ( [StringLiteral] ): move to the sub-key of that string.
   • For a BracketSegment ( [NumberLiteral] ): move to the array index NumberLiteral.
3. Stop once all segments have been resolved, returning the final value or indicating a missing path if any key/index does not exist.

---

8. Conclusion

This specification provides:

• Strict dot-notation compatibility (no leading/trailing/multiple consecutive dots).
• Flexible bracket notation to handle any string key (including spaces, symbols, or dots) or numeric index.
• Mixed usage that seamlessly allows dot segments for valid identifiers and bracket segments for anything else.