gsv

-module(gsv).
-compile([no_auto_import, nowarn_unused_vars, nowarn_unused_function, nowarn_nomatch, inline]).
-define(FILEPATH, "src/gsv.gleam").
-export([to_lists/2, to_dicts/2, from_lists/3, from_dicts/3]).
-export_type([line_ending/0, error/0, splitters/0]).

-if(?OTP_RELEASE >= 27).
-define(MODULEDOC(Str), -moduledoc(Str)).
-define(DOC(Str), -doc(Str)).
-else.
-define(MODULEDOC(Str), -compile([])).
-define(DOC(Str), -compile([])).
-endif.

-type line_ending() :: windows | unix.

-type error() :: {unescaped_quote, integer()} |
    {missing_closing_quote, integer()}.

-type splitters() :: {splitters,
        binary(),
        splitter:splitter(),
        splitter:splitter(),
        splitter:splitter()}.

-file("src/gsv.gleam", 57).
-spec line_ending_to_string(line_ending()) -> binary().
line_ending_to_string(Le) ->
    case Le of
        windows ->
            <<"\r\n"/utf8>>;

        unix ->
            <<"\n"/utf8>>
    end.

-file("src/gsv.gleam", 213).
?DOC(
    " This parses an escaped field from the start of the line, once we've already\n"
    " found the opening `\"`.\n"
    "\n"
    " > You might have noticed this also needs the rest of the document and just\n"
    " > the current line is not enough. This is because an escaped field might\n"
    " > contain newlines, and so we could end up needing to get more lines from\n"
    " > the `rest` of the csv to properly parse the row.\n"
).
-spec escaped_loop(
    binary(),
    binary(),
    binary(),
    splitters(),
    integer(),
    integer(),
    binary()
) -> {ok, {binary(), binary(), binary(), integer(), binary()}} |
    {error, error()}.
escaped_loop(Line, Newline, Rest, Splitters, Start, Line_number, Field) ->
    case splitter_ffi:split(erlang:element(5, Splitters), Line) of
        {Line_piece, <<""/utf8>>, <<""/utf8>>} ->
            case splitter_ffi:split(erlang:element(3, Splitters), Rest) of
                {<<""/utf8>>, <<""/utf8>>, <<""/utf8>>} ->
                    {error, {missing_closing_quote, Start}};

                {Line@1, Rest_newline, Rest@1} ->
                    Line@2 = <<<<Line_piece/binary, Newline/binary>>/binary,
                        Line@1/binary>>,
                    Line_number@1 = Line_number + 1,
                    escaped_loop(
                        Line@2,
                        Rest_newline,
                        Rest@1,
                        Splitters,
                        Start,
                        Line_number@1,
                        Field
                    )
            end;

        {Field_piece, <<"\"\""/utf8>>, Line@3} ->
            Field@1 = <<<<Field/binary, Field_piece/binary>>/binary, "\""/utf8>>,
            escaped_loop(
                Line@3,
                Newline,
                Rest,
                Splitters,
                Start,
                Line_number,
                Field@1
            );

        {Field_piece@1, _, <<""/utf8>>} ->
            Field@2 = case Field of
                <<""/utf8>> ->
                    Field_piece@1;

                _ ->
                    <<Field/binary, Field_piece@1/binary>>
            end,
            {ok, {<<""/utf8>>, Newline, Rest, Line_number, Field@2}};

        {Field_piece@2, _, Line@4} ->
            case gleam_stdlib:string_starts_with(
                Line@4,
                erlang:element(2, Splitters)
            ) of
                false ->
                    {error, {unescaped_quote, Line_number}};

                true ->
                    Line@5 = erlang:element(
                        3,
                        splitter_ffi:split(erlang:element(4, Splitters), Line@4)
                    ),
                    Field@3 = case Field of
                        <<""/utf8>> ->
                            Field_piece@2;

                        _ ->
                            <<Field/binary, Field_piece@2/binary>>
                    end,
                    {ok, {Line@5, Newline, Rest, Line_number, Field@3}}
            end
    end.

-file("src/gsv.gleam", 161).
?DOC(
    " This parses all the fields from the given line, returning them along with\n"
    " the reamining part of the csv that was not parsed.\n"
    "\n"
    " > You might have noticed that this doesn't just takes the line to parse as\n"
    " > input, but it also needs the rest of the csv file. It might sound counter\n"
    " > intuitive, but its need is explained in more detail in the `escape_loop`\n"
    " > function.\n"
).
-spec field_loop(
    binary(),
    binary(),
    binary(),
    splitters(),
    integer(),
    list(binary())
) -> {ok, {binary(), integer(), list(binary())}} | {error, error()}.
field_loop(Line, Newline, Rest, Splitters, Line_number, Fields) ->
    case splitter_ffi:split(erlang:element(4, Splitters), Line) of
        {Field, <<""/utf8>>, <<""/utf8>>} ->
            {ok, {Rest, Line_number, lists:reverse([Field | Fields])}};

        {<<""/utf8>>, <<"\""/utf8>>, Line@1} ->
            Start = Line_number,
            Field@1 = escaped_loop(
                Line@1,
                Newline,
                Rest,
                Splitters,
                Start,
                Line_number,
                <<""/utf8>>
            ),
            case Field@1 of
                {error, Error} ->
                    {error, Error};

                {ok, {<<""/utf8>>, _, Rest@1, Line_number@1, Field@2}} ->
                    {ok,
                        {Rest@1,
                            Line_number@1,
                            lists:reverse([Field@2 | Fields])}};

                {ok, {Line@2, Newline@1, Rest@2, Line_number@2, Field@3}} ->
                    Fields@1 = [Field@3 | Fields],
                    field_loop(
                        Line@2,
                        Newline@1,
                        Rest@2,
                        Splitters,
                        Line_number@2,
                        Fields@1
                    )
            end;

        {_, <<"\""/utf8>>, _} ->
            {error, {unescaped_quote, Line_number}};

        {Field@4, _, Line@3} ->
            Fields@2 = [Field@4 | Fields],
            field_loop(Line@3, Newline, Rest, Splitters, Line_number, Fields@2)
    end.

-file("src/gsv.gleam", 135).
-spec line_loop(binary(), splitters(), integer()) -> {ok,
        {binary(), integer(), list(binary())}} |
    {error, error()}.
line_loop(Csv, Splitters, Line_number) ->
    case splitter_ffi:split(erlang:element(3, Splitters), Csv) of
        {<<""/utf8>>, _, <<""/utf8>>} ->
            {ok, {<<""/utf8>>, Line_number, []}};

        {<<""/utf8>>, _, Rest} ->
            line_loop(Rest, Splitters, Line_number + 1);

        {Line, Newline, Rest@1} ->
            field_loop(Line, Newline, Rest@1, Splitters, Line_number + 1, [])
    end.

-file("src/gsv.gleam", 118).
-spec lines_loop(binary(), splitters(), integer(), list(list(binary()))) -> {ok,
        list(list(binary()))} |
    {error, error()}.
lines_loop(Csv, Splitters, Line_number, Lines) ->
    case Csv of
        <<""/utf8>> ->
            {ok, lists:reverse(Lines)};

        _ ->
            case line_loop(Csv, Splitters, Line_number) of
                {error, Error} ->
                    {error, Error};

                {ok, {Rest, Line_number@1, Line}} ->
                    lines_loop(Rest, Splitters, Line_number@1, [Line | Lines])
            end
    end.

-file("src/gsv.gleam", 103).
?DOC(
    " Parses a csv string into a list of lists of strings: each line of the csv\n"
    " will be turned into a list with an item for each field.\n"
    "\n"
    " ## Examples\n"
    "\n"
    " ```gleam\n"
    " \"hello, world\n"
    " goodbye, mars\"\n"
    " |> gsv.to_lists(seperator: \",\")\n"
    " // Ok([\n"
    " //    [\"hello\", \" world\"],\n"
    " //    [\"goodbye\", \" mars\"],\n"
    " // ])\n"
    " ```\n"
    "\n"
    " > This implementation tries to stick as closely as possible to\n"
    " > [RFC4180](https://www.ietf.org/rfc/rfc4180.txt), with a couple of notable\n"
    " > differences:\n"
    " > - both `\\n` and `\\r\\n` line endings are accepted.\n"
    " > - empty lines are allowed and just ignored.\n"
    " > - lines are not forced to all have the same number of fields.\n"
    " > - a line can start with an empty field `,two,three`.\n"
    " > - a line can end with an empty field `one,two,`.\n"
).
-spec to_lists(binary(), binary()) -> {ok, list(list(binary()))} |
    {error, error()}.
to_lists(Csv, Separator) ->
    Splitters = {splitters,
        Separator,
        splitter:new([<<"\r\n"/utf8>>, <<"\n"/utf8>>]),
        splitter:new([Separator, <<"\""/utf8>>]),
        splitter:new([<<"\"\""/utf8>>, <<"\""/utf8>>])},
    lines_loop(Csv, Splitters, 0, []).

-file("src/gsv.gleam", 318).
?DOC(
    " Parses a csv string into a list of dicts: the first line of the csv is\n"
    " interpreted as the headers' row and each of the following lines is turned\n"
    " into a dict with a value for each of the headers.\n"
    "\n"
    " If a field is empty then it won't be added to the dict.\n"
    "\n"
    " ## Examples\n"
    "\n"
    " ```gleam\n"
    " \"pet,name,cuteness\n"
    " dog,Fido,100\n"
    " cat,,1000\n"
    " \"\n"
    " |> gsv.to_dicts(separator: \",\")\n"
    " // Ok([\n"
    " //    dict.from_list([\n"
    " //      #(\"pet\", \"dog\"), #(\"name\", \"Fido\"), #(\"cuteness\", \"100\")\n"
    " //    ]),\n"
    " //    dict.from_list([\n"
    " //      #(\"pet\", \"cat\"), #(\"cuteness\", \"1000\")\n"
    " //    ]),\n"
    " // ])\n"
    " ```\n"
    "\n"
    " > Just list `to_lists` this implementation tries to stick as closely as\n"
    " > possible to [RFC4180](https://www.ietf.org/rfc/rfc4180.txt).\n"
    " > You can look at `to_lists`' documentation to see how it differs from the\n"
    " > RFC.\n"
).
-spec to_dicts(binary(), binary()) -> {ok,
        list(gleam@dict:dict(binary(), binary()))} |
    {error, error()}.
to_dicts(Input, Field_separator) ->
    gleam@result:map(to_lists(Input, Field_separator), fun(Rows) -> case Rows of
                [] ->
                    [];

                [Headers | Rows@1] ->
                    Headers@1 = erlang:list_to_tuple(Headers),
                    gleam@list:map(
                        Rows@1,
                        fun(Row) ->
                            gleam@list:index_fold(
                                Row,
                                maps:new(),
                                fun(Row@1, Field, Index) -> case Field of
                                        <<""/utf8>> ->
                                            Row@1;

                                        _ ->
                                            case glearray_ffi:get(
                                                Headers@1,
                                                Index
                                            ) of
                                                {ok, Header} ->
                                                    gleam@dict:insert(
                                                        Row@1,
                                                        Header,
                                                        Field
                                                    );

                                                {error, _} ->
                                                    Row@1
                                            end
                                    end end
                            )
                        end
                    )
            end end).

-file("src/gsv.gleam", 377).
-spec escape_field(binary(), binary()) -> binary().
escape_field(Field, Separator) ->
    case gleam_stdlib:contains_string(Field, <<"\""/utf8>>) of
        true ->
            <<<<"\""/utf8,
                    (gleam@string:replace(Field, <<"\""/utf8>>, <<"\"\""/utf8>>))/binary>>/binary,
                "\""/utf8>>;

        false ->
            case gleam_stdlib:contains_string(Field, Separator) orelse gleam_stdlib:contains_string(
                Field,
                <<"\n"/utf8>>
            ) of
                true ->
                    <<<<"\""/utf8, Field/binary>>/binary, "\""/utf8>>;

                false ->
                    Field
            end
    end.

-file("src/gsv.gleam", 362).
?DOC(
    " Takes a list of lists of strings and turns it to a csv string, automatically\n"
    " escaping all fields that contain double quotes or line endings.\n"
    "\n"
    " ## Examples\n"
    "\n"
    " ```gleam\n"
    " let rows = [[\"hello\", \"world\"], [\"goodbye\", \"mars\"]]\n"
    " from_lists(rows, separator: \",\", line_ending: Unix)\n"
    " // \"hello,world\n"
    " // goodbye,mars\"\n"
    " ```\n"
    "\n"
    " ```gleam\n"
    " let rows = [[]]\n"
    " ```\n"
).
-spec from_lists(list(list(binary())), binary(), line_ending()) -> binary().
from_lists(Rows, Separator, Line_ending) ->
    Line_ending@1 = line_ending_to_string(Line_ending),
    _pipe@1 = gleam@list:map(
        Rows,
        fun(Row) ->
            _pipe = gleam@list:map(
                Row,
                fun(_capture) -> escape_field(_capture, Separator) end
            ),
            gleam@string:join(_pipe, Separator)
        end
    ),
    _pipe@2 = gleam@string:join(_pipe@1, Line_ending@1),
    gleam@string:append(_pipe@2, Line_ending@1).

-file("src/gsv.gleam", 414).
-spec row_dict_to_list(gleam@dict:dict(binary(), binary()), list(binary())) -> list(binary()).
row_dict_to_list(Row, Headers) ->
    gleam@list:map(
        Headers,
        fun(Header) -> case gleam_stdlib:map_get(Row, Header) of
                {ok, Field} ->
                    Field;

                {error, nil} ->
                    <<""/utf8>>
            end end
    ).

-file("src/gsv.gleam", 394).
?DOC(
    " Takes a list of dicts and writes it to a csv string.\n"
    " Will automatically escape strings that contain double quotes or\n"
    " line endings with double quotes (in csv, double quotes get escaped by doing\n"
    " a double doublequote)\n"
    " The string `he\"llo\\n` becomes `\"he\"\"llo\\n\"`\n"
).
-spec from_dicts(
    list(gleam@dict:dict(binary(), binary())),
    binary(),
    line_ending()
) -> binary().
from_dicts(Rows, Separator, Line_ending) ->
    case Rows of
        [] ->
            <<""/utf8>>;

        _ ->
            Headers = begin
                _pipe = Rows,
                _pipe@1 = gleam@list:flat_map(_pipe, fun maps:keys/1),
                _pipe@2 = gleam@list:unique(_pipe@1),
                gleam@list:sort(_pipe@2, fun gleam@string:compare/2)
            end,
            Rows@1 = gleam@list:map(
                Rows,
                fun(_capture) -> row_dict_to_list(_capture, Headers) end
            ),
            from_lists([Headers | Rows@1], Separator, Line_ending)
    end.