Files
Odin/core/encoding/csv/reader.odin

488 lines
13 KiB
Odin

// package csv reads and writes comma-separated values (CSV) files.
// This package supports the format described in [[ RFC 4180; https://tools.ietf.org/html/rfc4180.html ]]
package encoding_csv
import "core:bufio"
import "core:bytes"
import "core:io"
import "core:strings"
import "core:unicode/utf8"
// Reader is a data structure used for reading records from a CSV-encoded file
//
// The associated procedures for Reader expects its input to conform to RFC 4180.
Reader :: struct {
// comma is the field delimiter
// reader_init will set it to be ','
// A "comma" must be a valid rune, nor can it be \r, \n, or the Unicode replacement character (0xfffd)
comma: rune,
// comment, if not 0, is the comment character
// Lines beginning with the comment character without a preceding whitespace are ignored
comment: rune,
// fields_per_record is the number of expected fields per record
// if fields_per_record is >0, 'read' requires each record to have that field count
// if fields_per_record is 0, 'read' sets it to the field count in the first record
// if fields_per_record is <0, no check is made and records may have a variable field count
fields_per_record: int,
// If trim_leading_space is true, leading whitespace in a field is ignored
// This is done even if the field delimiter (comma), is whitespace
trim_leading_space: bool,
// If lazy_quotes is true, a quote may appear in an unquoted field and a non-doubled quote may appear in a quoted field
lazy_quotes: bool,
// multiline_fields, when set to true, will treat a field starting with a " as a multiline string
// therefore, instead of reading until the next \n, it'll read until the next "
multiline_fields: bool,
// reuse_record controls whether calls to 'read' may return a slice using the backing buffer
// for performance
// By default, each call to 'read' returns a newly allocated slice
reuse_record: bool,
// reuse_record_buffer controls whether calls to 'read' clone the strings of each field or uses
// the data stored in record buffer for performance
// By default, each call to 'read' clones the strings of each field
reuse_record_buffer: bool,
// internal buffers
r: bufio.Reader,
line_count: int, // current line being read in the CSV file
raw_buffer: [dynamic]byte,
record_buffer: [dynamic]byte,
field_indices: [dynamic]int,
last_record: [dynamic]string,
sr: strings.Reader, // used by reader_init_with_string
// Set and used by the iterator. Query using `iterator_last_error`
last_iterator_error: Error,
}
Reader_Error_Kind :: enum {
Bare_Quote,
Quote,
Field_Count,
Invalid_Delim,
}
reader_error_kind_string := [Reader_Error_Kind]string{
.Bare_Quote = "bare \" in non-quoted field",
.Quote = "extra or missing \" in quoted field",
.Field_Count = "wrong field count",
.Invalid_Delim = "invalid delimiter",
}
Reader_Error :: struct {
kind: Reader_Error_Kind,
start_line: int,
line: int,
column: int,
expected, got: int, // used by .Field_Count
}
Error :: union {
Reader_Error,
io.Error,
}
DEFAULT_RECORD_BUFFER_CAPACITY :: 256
// reader_init initializes a new Reader from r
reader_init :: proc(reader: ^Reader, r: io.Reader, buffer_allocator := context.allocator) {
switch reader.comma {
case '\x00', '\n', '\r', 0xfffd:
reader.comma = ','
}
context.allocator = buffer_allocator
reserve(&reader.record_buffer, DEFAULT_RECORD_BUFFER_CAPACITY)
reserve(&reader.raw_buffer, 0)
reserve(&reader.field_indices, 0)
reserve(&reader.last_record, 0)
bufio.reader_init(&reader.r, r)
}
// reader_init_with_string initializes a new Reader from s
reader_init_with_string :: proc(reader: ^Reader, s: string, buffer_allocator := context.allocator) {
strings.reader_init(&reader.sr, s)
r, _ := io.to_reader(strings.reader_to_stream(&reader.sr))
reader_init(reader, r, buffer_allocator)
}
// reader_destroy destroys a Reader
reader_destroy :: proc(r: ^Reader) {
delete(r.raw_buffer)
delete(r.record_buffer)
delete(r.field_indices)
delete(r.last_record)
bufio.reader_destroy(&r.r)
}
/*
Returns a record at a time.
for record, row_idx in csv.iterator_next(&r) { ... }
TIP: If you process the results within the loop and don't need to own the results,
you can set the Reader's `reuse_record` and `reuse_record_buffer` to true;
you won't need to delete the record or its fields.
*/
iterator_next :: proc(r: ^Reader) -> (record: []string, idx: int, err: Error, more: bool) {
record, r.last_iterator_error = read(r)
return record, r.line_count - 1, r.last_iterator_error, r.last_iterator_error == nil
}
// Get last CSV parse error if we ignored it in the iterator loop
//
// for record, row_idx in csv.iterator_next(&r) { ... }
iterator_last_error :: proc(r: Reader) -> (err: Error) {
return r.last_iterator_error
}
// read reads a single record (a slice of fields) from r
//
// All \r\n sequences are normalized to \n, including multi-line field
@(require_results)
read :: proc(r: ^Reader, allocator := context.allocator) -> (record: []string, err: Error) {
if r.reuse_record {
record, err = _read_record(r, &r.last_record, allocator)
resize(&r.last_record, len(record))
copy(r.last_record[:], record)
} else {
record, err = _read_record(r, nil, allocator)
}
return
}
// is_io_error checks where an Error is a specific io.Error kind
@(require_results)
is_io_error :: proc(err: Error, io_err: io.Error) -> bool {
if v, ok := err.(io.Error); ok {
return v == io_err
}
return false
}
// read_all reads all the remaining records from r.
// Each record is a slice of fields.
// read_all is defined to read until an EOF, and does not treat EOF as an error
@(require_results)
read_all :: proc(r: ^Reader, allocator := context.allocator) -> ([][]string, Error) {
context.allocator = allocator
records: [dynamic][]string
for {
record, rerr := _read_record(r, nil, allocator)
if is_io_error(rerr, .EOF) {
return records[:], nil
}
if rerr != nil {
// allow for a partial read
if record != nil {
append(&records, record)
}
return records[:], rerr
}
append(&records, record)
}
}
// read reads a single record (a slice of fields) from the provided input.
@(require_results)
read_from_string :: proc(input: string, record_allocator := context.allocator, buffer_allocator := context.allocator) -> (record: []string, n: int, err: Error) {
ir: strings.Reader
strings.reader_init(&ir, input)
input_reader, _ := io.to_reader(strings.reader_to_stream(&ir))
r: Reader
reader_init(&r, input_reader, buffer_allocator)
defer reader_destroy(&r)
record, err = read(&r, record_allocator)
n = int(r.r.r)
return
}
// read_all reads all the remaining records from the provided input.
@(require_results)
read_all_from_string :: proc(input: string, records_allocator := context.allocator, buffer_allocator := context.allocator) -> ([][]string, Error) {
ir: strings.Reader
strings.reader_init(&ir, input)
input_reader, _ := io.to_reader(strings.reader_to_stream(&ir))
r: Reader
reader_init(&r, input_reader, buffer_allocator)
defer reader_destroy(&r)
return read_all(&r, records_allocator)
}
@(private, require_results)
is_valid_delim :: proc(r: rune) -> bool {
switch r {
case 0, '"', '\r', '\n', utf8.RUNE_ERROR:
return false
}
return utf8.valid_rune(r)
}
@(private, require_results)
_read_record :: proc(r: ^Reader, dst: ^[dynamic]string, allocator := context.allocator) -> ([]string, Error) {
@(require_results)
read_line :: proc(r: ^Reader) -> ([]byte, io.Error) {
if !r.multiline_fields {
line, err := bufio.reader_read_slice(&r.r, '\n')
if err == .Buffer_Full {
clear(&r.raw_buffer)
append(&r.raw_buffer, ..line)
for err == .Buffer_Full {
line, err = bufio.reader_read_slice(&r.r, '\n')
append(&r.raw_buffer, ..line)
}
line = r.raw_buffer[:]
}
if len(line) > 0 && err == .EOF {
err = nil
if line[len(line)-1] == '\r' {
line = line[:len(line)-1]
}
}
r.line_count += 1
// normalize \r\n to \n
n := len(line)
for n >= 2 && string(line[n-2:]) == "\r\n" {
line[n-2] = '\n'
line = line[:n-1]
}
return line, err
} else {
// Reading a "line" that can possibly contain multiline fields.
// Unfortunately, this means we need to read a character at a time.
err: io.Error
cur: rune
is_quoted: bool
field_length := 0
clear(&r.raw_buffer)
read_loop: for err == .None {
cur, _, err = bufio.reader_read_rune(&r.r)
if err != .None { break read_loop }
switch cur {
case '"':
is_quoted = field_length == 0
field_length += 1
case '\n', '\r':
is_quoted or_break read_loop
case r.comma:
field_length = 0
case:
field_length += 1
}
rune_buf, rune_len := utf8.encode_rune(cur)
append(&r.raw_buffer, ..rune_buf[:rune_len])
}
return r.raw_buffer[:], err
}
unreachable()
}
@(require_results)
length_newline :: proc(b: []byte) -> int {
if len(b) > 0 && b[len(b)-1] == '\n' {
return 1
}
return 0
}
@(require_results)
next_rune :: proc(b: []byte) -> rune {
r, _ := utf8.decode_rune(b)
return r
}
if r.comma == r.comment ||
!is_valid_delim(r.comma) ||
(r.comment != 0 && !is_valid_delim(r.comment)) {
err := Reader_Error{
kind = .Invalid_Delim,
line = r.line_count,
}
return nil, err
}
line, full_line: []byte
err_read: io.Error
for err_read == nil {
line, err_read = read_line(r)
if r.comment != 0 && next_rune(line) == r.comment {
line = nil
continue
}
if err_read == nil && len(line) == length_newline(line) {
line = nil
continue
}
full_line = line
break
}
if is_io_error(err_read, .EOF) {
return nil, err_read
}
err: Error
quote_len :: len(`"`)
comma_len := utf8.rune_size(r.comma)
record_line := r.line_count
clear(&r.record_buffer)
clear(&r.field_indices)
parse_field: for {
if r.trim_leading_space {
line = bytes.trim_left_space(line)
}
if len(line) == 0 || line[0] != '"' {
i := bytes.index_rune(line, r.comma)
field := line
if i >= 0 {
field = field[:i]
} else {
field = field[:len(field) - length_newline(field)]
}
if !r.lazy_quotes {
if j := bytes.index_byte(field, '"'); j >= 0 {
column := utf8.rune_count(full_line[:len(full_line) - len(line[j:])])
err = Reader_Error{
kind = .Bare_Quote,
start_line = record_line,
line = r.line_count,
column = column,
}
break parse_field
}
}
append(&r.record_buffer, ..field)
append(&r.field_indices, len(r.record_buffer))
if i >= 0 {
line = line[i+comma_len:]
continue parse_field
}
break parse_field
} else {
line = line[quote_len:]
for {
i := bytes.index_byte(line, '"')
switch {
case i >= 0:
append(&r.record_buffer, ..line[:i])
line = line[i+quote_len:]
switch ch := next_rune(line); {
case ch == '"': // append quote
append(&r.record_buffer, '"')
line = line[quote_len:]
case ch == r.comma: // end of field
line = line[comma_len:]
append(&r.field_indices, len(r.record_buffer))
continue parse_field
case length_newline(line) == len(line): // end of line
append(&r.field_indices, len(r.record_buffer))
break parse_field
case r.lazy_quotes: // bare quote
append(&r.record_buffer, '"')
case: // invalid non-escaped quote
column := utf8.rune_count(full_line[:len(full_line) - len(line) - quote_len])
err = Reader_Error{
kind = .Quote,
start_line = record_line,
line = r.line_count,
column = column,
}
break parse_field
}
case len(line) > 0:
append(&r.record_buffer, ..line)
if err_read != nil {
break parse_field
}
line, err_read = read_line(r)
if is_io_error(err_read, .EOF) {
err_read = nil
}
full_line = line
case:
if !r.lazy_quotes && err_read == nil {
column := utf8.rune_count(full_line)
err = Reader_Error{
kind = .Quote,
start_line = record_line,
line = r.line_count,
column = column,
}
break parse_field
}
append(&r.field_indices, len(r.record_buffer))
break parse_field
}
}
}
}
if err == nil && err_read != nil {
err = err_read
}
context.allocator = allocator
dst := dst
str := string(r.record_buffer[:])
if dst == nil {
// use local variable
dst = &([dynamic]string){}
}
clear(dst)
resize(dst, len(r.field_indices))
pre_idx: int
for idx, i in r.field_indices {
field := str[pre_idx:idx]
if !r.reuse_record_buffer {
field = strings.clone(field)
}
dst[i] = field
pre_idx = idx
}
if r.fields_per_record > 0 {
if len(dst) != r.fields_per_record && err == nil {
err = Reader_Error{
kind = .Field_Count,
start_line = record_line,
line = r.line_count,
expected = r.fields_per_record,
got = len(dst),
}
}
} else if r.fields_per_record == 0 {
r.fields_per_record = len(dst)
}
return dst[:], err
}