Package 'naxp'

Check text against a naxp

Description

naxp_accepts() says whether a naxp accepts each text, and naxp_canonical_form() gives the form each accepted text decodes to.

Usage

naxp_accepts(text, naxp)

naxp_canonical_form(text, naxp)

Arguments

text	A character vector.
naxp	A naxp from naxp_parse(), a pattern string or a naxp_encoded vector, whose naxp is used. For naxp_decode(), it is needed only where x is not a naxp_encoded vector.

Details

Neither needs the text's encoded value, so both work for a naxp of any size.

Value

• naxp_accepts(): a logical vector the same length as text, with its names. It is FALSE, never NA, for NA and for invalid text.

• naxp_canonical_form(): a character vector the same length as text, with its names, holding NA for NA and for invalid text.

Examples

nl_postcode <- naxp_parse(r"(\9\9\9\9 \s!! \A\A)")

naxp_accepts(c("1012AB", "1012 AB", "1012-AB"), nl_postcode)
naxp_canonical_form(c("1012AB", "1012 AB", "1012-AB"), nl_postcode)

---

Compare two naxps

Description

naxp_compare() says how b stands to a: what happens to text and encoded values held under a if b replaces it. naxp_first_divergent_value() finds the lowest encoded value the two decode differently.

Usage

naxp_compare(a, b, budget = NULL)

naxp_first_divergent_value(a, b)

Arguments

a	The naxp the data was encoded with: a naxp from naxp_parse(), a pattern string or a naxp_encoded vector, whose naxp is used.
b	The naxp proposed to replace it, likewise.
budget	NULL for the library's default budget of 200 000 product states, or a positive whole number.

Details

naxp_compare() gives three set relationships, each from a's point of view:

• accepted_text: how the strings a accepts stand to those b accepts.

• encoding: how the (text, encoded value) pairs a defines stand to those b defines. Replacing a with b is safe for stored values exactly when this is "equal" or "subset_of".

• canonical_forms: how the canonical forms a prints stand to those b prints.

Each is "equal", "subset_of", "superset_of" or "incomparable".

The encoding relationship is decided by walking the two naxps together, which has a budget. No naxp anybody has reason to write comes near the default, but a pair that does cannot be decided, and naxp_compare() then raises an error rather than guess.

Value

• naxp_compare(): a named character vector with elements accepted_text, encoding and canonical_forms.

• naxp_first_divergent_value(): a single encoded value, or 0 where every value both naxps hold decodes alike. Values only one naxp holds do not count. It is an integer where both naxps' encoded values are integers, and otherwise a double.

Examples

with_space <- naxp_parse(r"(\9\9\9\9 \s!! \A\A)")
without_space <- naxp_parse(r"(\9\9\9\9 \s!? \A\A)")

naxp_compare(with_space, without_space)
naxp_first_divergent_value(with_space, without_space)

---

Emit a naxp as source code

Description

Writes a naxp as source code in another language: a fragment of declarations that answer the same questions this package does, for that one naxp, with no dependency on any naxp library. The header comment, includes, namespace or wrapping class around it are yours to add.

Usage

naxp_emit(
  naxp,
  language = c("csharp", "javascript", "c", "cpp"),
  prefix = "",
  value_type = c("uint64", "int64", "uint32", "int32", "uint16", "int16", "uint8",
    "int8"),
  initial_indent = "",
  new_line = "\n",
  indent = "\t"
)

Arguments

naxp	A naxp from naxp_parse(), a pattern string or a naxp_encoded vector, whose naxp is used.
language	The language to emit: "csharp", "javascript", "c" (C99) or "cpp" (C++17).
prefix	What every generated name starts with, so several naxps can share one scope. "" gives the bare names. Each language applies its own casing convention to it.
value_type	The integer type the generated code uses for encoded values. The naxp's maximum encoded value must fit it.
initial_indent	What every line starts with ahead of its own indentation, so the fragment can sit inside an indented wrapper.
new_line	What ends every line.
indent	What one level of indentation is written as.

Value

A single string: the fragment.

Examples

cat(naxp_emit(r"(\9\9\9\9 \s!! \A\A)", "javascript", prefix = "nl_postcode"))

---

Encode and decode text

Description

naxp_encode() encodes text as encoded values under a naxp, and naxp_decode() decodes them back to the canonical form of the text.

Usage

naxp_encode(text, naxp)

naxp_decode(x, naxp = NULL)

Arguments

text	A character vector.
naxp	A naxp from naxp_parse(), a pattern string or a naxp_encoded vector, whose naxp is used. For naxp_decode(), it is needed only where x is not a naxp_encoded vector.
x	A naxp_encoded vector, or an integer or double vector of encoded values. Bare values must be whole numbers from 0 to the naxp's maximum encoded value, or NA.

Details

Each text a naxp accepts encodes to a whole number from 1 to its maximum encoded value. Several texts may share a value, such as a postcode written with and without a space, and each value decodes to one canonical form.

Text the naxp does not accept is invalid and encodes to 0, as in every naxp library. So does NA, and so does any text that is not ASCII, whatever its marked encoding. Encoded values are never NA. naxp_decode() returns NA for 0, because no other string can stand for it: a naxp may accept the empty string.

Encoded values are held as integers where the naxp's maximum encoded value is at most 2^31 - 1, and as doubles where it is at most 2^53. Above 2^53, which is the largest whole number R holds exactly, both functions raise an error.

Value

• naxp_encode(): a naxp_encoded vector the same length as text, with its names.

• naxp_decode(): a character vector the same length as x, with its names, holding NA where a value is 0 or NA.

See Also

naxp_accepts() and naxp_canonical_form() to check text without encoding it.

Examples

nl_postcode <- naxp_parse(r"(\9\9\9\9 \s!! \A\A)")

codes <- naxp_encode(c("1012AB", "1012 AB", "not a postcode"), nl_postcode)
codes
as.integer(codes)

naxp_decode(codes)
naxp_decode(684114L, nl_postcode)

---

Encoded values

Description

A naxp_encoded vector holds encoded values together with the naxp that encoded them, so it can always be decoded. naxp_encode() returns one, and naxp_encoded() makes one from bare values, such as values read from a database or a file.

Usage

naxp_encoded(values, naxp)

Arguments

values	An integer or double vector of whole numbers from 0 to the naxp's maximum encoded value. NA is an error.
naxp	A naxp from naxp_parse(), a pattern string or a naxp_encoded vector, whose naxp is used.

Details

The values are integers where the naxp's maximum encoded value is at most 2^31 - 1, and doubles where it is at most 2^53. Above 2^53 is an error. 0 stands for text that is not valid. Encoded values are never NA.

A naxp_encoded vector prints as the text it decodes to, with 0 shown as ⁠<invalid (0)>⁠. as.character() gives the text, with NA for 0, and as.integer() or as.numeric() give the bare values.

Subsetting, rep(), rev(), unique() and sort() return naxp_encoded vectors with the same naxp, and split() returns a list of them. A naxp_encoded vector can be a data frame column. order() and match() return integer indices, and table() counts each distinct value, labelled with its text, excluding 0. Assigning into a naxp_encoded vector accepts another with the same naxp, or text, which is encoded. c() combines naxp_encoded vectors with the same naxp.

Encoded values are compared only for equality. == and != compare a naxp_encoded vector with one with the same naxp, with text, which is encoded first, or with numbers as bare values, so x == 0 finds the invalid text. Arithmetic, <, > and summaries such as max() raise errors. sort() and order() sort by encoded value.

Value

A naxp_encoded vector the same length as values, with its names.

Examples

nl_postcode <- naxp_parse(r"(\9\9\9\9 \s!! \A\A)")

codes <- naxp_encoded(c(684114, 0, 6760000), nl_postcode)
codes
codes == "1012 AB"
codes == 0

data.frame(postcode = codes)

---

Parse a naxp

Description

Parses the pattern of a naxp, checking that it is valid, and returns a naxp for the other functions in the package.

Usage

naxp_parse(pattern)

Arguments

pattern

A single string: the pattern of the naxp.

Details

A naxp holds only its pattern. The compiled naxp is kept for the rest of the session, so parsing the same pattern again is cheap, and a naxp saved with saveRDS() or sent to a parallel worker is compiled again when first used.

Every function that takes a naxp also accepts a pattern string, which it parses in the same way, or a naxp_encoded vector, whose naxp it uses.

Value

A naxp: an object of class naxp holding the pattern.

Errors

An invalid pattern raises a condition of class naxp_error, which is also an error. Its fields are named as in the other naxp libraries:

• code: the error code, such as "NAXP1025", which is the same in every version.

• message: what is wrong, and where practical what to write instead.

• offset: where the error starts, in characters from the start of the pattern, counting from 0 as the other naxp libraries do.

• length: how many characters the error covers.

• pattern: the pattern.

Each error code is explained at https://naxp.org/error-codes/.

See Also

naxp_encode() to encode text, and naxp_pattern() for a naxp's properties.

Examples

nl_postcode <- naxp_parse(r"(\9\9\9\9 \s!! \A\A)")
nl_postcode

# An invalid pattern raises a naxp_error
tryCatch(
  naxp_parse(r"(\9\9\9\9 [A-)"),
  naxp_error = function(error) error$code
)

---

Properties of a naxp

Description

The pattern a naxp was parsed from, its largest encoded value and the length of its longest canonical form.

Usage

naxp_pattern(naxp)

naxp_max_encoded_value(naxp)

naxp_max_length(naxp)

Arguments

naxp	A naxp from naxp_parse(), a pattern string or a naxp_encoded vector, whose naxp is used.

Details

Encoded values run from 1 to the maximum with no gaps, so the maximum is also the number of distinct encoded values. It is an integer where it is at most 2^31 - 1 and a double where it is at most 2^53. Above 2^53, which is the largest whole number R holds exactly, naxp_max_encoded_value() raises an error, although printing the naxp still shows the maximum.

Value

• naxp_pattern(): a single string.

• naxp_max_encoded_value(): a single integer or double.

• naxp_max_length(): a single integer.

Examples

nl_postcode <- naxp_parse(r"(\9\9\9\9 \s!! \A\A)")
naxp_pattern(nl_postcode)
naxp_max_encoded_value(nl_postcode)
naxp_max_length(nl_postcode)

# An encoded vector stands for the naxp that encoded it
codes <- naxp_encode(c("1012AB", "9999 ZZ"), nl_postcode)
naxp_pattern(codes)

---

Standard naxps

Description

naxp_standard() gives one of the standard naxps built into every naxp library, and naxp_standards() lists them all.

Usage

naxp_standard(id, variant = "default")

naxp_standards()

Arguments

id	A single string: the standard's id, such as "uk-postcode".
variant	A single string: the variant's id. "default" or "" gives the default.

Details

The standard naxps are for widely used code systems such as postcodes. Each has an id, such as "uk-postcode", and one or more variants, one of which is "default". The id and the variant are matched without regard to ASCII case. Once published, a standard naxp never changes, so values encoded with it keep their meaning.

The standard naxps are described at https://naxp.org/standard/.

Value

• naxp_standard(): a naxp.

• naxp_standards(): a data frame with one row per variant, each standard's default first, and character columns id, name (for display), variant, pattern and notes (what the variant accepts and prints, in Markdown).

Examples

uk_postcode <- naxp_standard("uk-postcode")
naxp_encode(c("EC4M 7AA", "EC4M7AA", "B1 1AA"), uk_postcode)

naxp_standard("nl-postcode", "no-separator")

naxp_standards()[, c("id", "variant")]

---