commit 790aeb70cd94c1d7215369e2de9cf57a8b9efaae · comet.sh/atex

CHANGELOG.md

···

       6
       6
        
       and this project adheres to

     

       7
       7
        
       [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

     

       8
       8
        
       

     

       9
       9
       +
       ## [Unreleased]

     

       10
       10
       +
       

     

       11
       11
       +
       ## Added

     

       12
       12
       +
       

     

       13
       13
       +
       - `Atex.TID` module for manipulating ATProto TIDs.

     

       14
       14
       +
       - `Atex.Base32Sortable` module for encoding/decoding numbers as

     

       15
       15
       +
         `base32-sortable` strings.

     

       16
       16
       +
       

     

       9
       17
        
       ## [0.1.0] - 2025-06-07

     

       10
       18
        
       

     

       11
       19
        
       Initial release.

+4 -3

lib/aturi.ex

···

       1
       1
        
       defmodule Atex.AtURI do

     

       2
       2
        
         @moduledoc """

     

       3
       3
        
         Struct and helper functions for manipulating `at://` URIs, which identify

     

       4
       4
       -
         specific records within the AT Protocol. For more information on the URI

     

       5
       5
       -
         scheme, refer to the ATProto spec: https://atproto.com/specs/at-uri-scheme.

     

       4
       4
       +
         specific records within the AT Protocol.

     

       5
       5
       +
       

     

       6
       6
       +
         ATProto spec: https://atproto.com/specs/at-uri-scheme

     

       6
       7
        
       

     

       7
       8
        
         This module only supports the restricted URI syntax used for the Lexicon

     

       8
       9
        
         `at-uri` type, with no support for query strings or fragments. If/when the

     
···

       154
       155
        
       end

     

       155
       156
        
       

     

       156
       157
        
       defimpl String.Chars, for: Atex.AtURI do

     

       157
       157
       -
         def to_string(%Atex.AtURI{} = uri), do: Atex.AtURI.to_string(uri)

     

       158
       158
       +
         def to_string(uri), do: Atex.AtURI.to_string(uri)

     

       158
       159
        
       end

+39

lib/base32_sortable.ex

···

       1
       1
       +
       defmodule Atex.Base32Sortable do

     

       2
       2
       +
         @moduledoc """

     

       3
       3
       +
         Codec for the base32-sortable encoding.

     

       4
       4
       +
         """

     

       5
       5
       +
       

     

       6
       6
       +
         @alphabet ~c(234567abcdefghijklmnopqrstuvwxyz)

     

       7
       7
       +
         @alphabet_len length(@alphabet)

     

       8
       8
       +
       

     

       9
       9
       +
         @doc """

     

       10
       10
       +
         Encode an integer as a base32-sortable string.

     

       11
       11
       +
         """

     

       12
       12
       +
         @spec encode(integer()) :: String.t()

     

       13
       13
       +
         def encode(int) when is_integer(int), do: do_encode(int, "")

     

       14
       14
       +
       

     

       15
       15
       +
         @spec do_encode(integer(), String.t()) :: String.t()

     

       16
       16
       +
         defp do_encode(0, acc), do: acc

     

       17
       17
       +
       

     

       18
       18
       +
         defp do_encode(int, acc) do

     

       19
       19
       +
           char_index = rem(int, @alphabet_len)

     

       20
       20
       +
           new_int = div(int, @alphabet_len)

     

       21
       21
       +
       

     

       22
       22
       +
           # Chars are prepended to the accumulator because rem/div is pulling them off the tail of the integer.

     

       23
       23
       +
           do_encode(new_int, <<Enum.at(@alphabet, char_index)>> <> acc)

     

       24
       24
       +
         end

     

       25
       25
       +
       

     

       26
       26
       +
         @doc """

     

       27
       27
       +
         Decode a base32-sortable string to an integer.

     

       28
       28
       +
         """

     

       29
       29
       +
         @spec decode(String.t()) :: integer()

     

       30
       30
       +
         def decode(str) when is_binary(str), do: do_decode(str, 0)

     

       31
       31
       +
       

     

       32
       32
       +
         @spec do_decode(String.t(), integer()) :: integer()

     

       33
       33
       +
         defp do_decode(<<>>, acc), do: acc

     

       34
       34
       +
       

     

       35
       35
       +
         defp do_decode(<<char::utf8, rest::binary>>, acc) do

     

       36
       36
       +
           i = Enum.find_index(@alphabet, fn x -> x == char end)

     

       37
       37
       +
           do_decode(rest, acc * @alphabet_len + i)

     

       38
       38
       +
         end

     

       39
       39
       +
       end

+169

lib/tid.ex

···

       1
       1
       +
       defmodule Atex.TID do

     

       2
       2
       +
         @moduledoc """

     

       3
       3
       +
         Struct and helper functions for dealing with AT Protocol TIDs (Timestamp

     

       4
       4
       +
         Identifiers), a 13-character string representation of a 64-bit number

     

       5
       5
       +
         comprised of a Unix timestamp (in microsecond precision) and a random "clock

     

       6
       6
       +
         identifier" to help avoid collisions.

     

       7
       7
       +
       

     

       8
       8
       +
         ATProto spec: https://atproto.com/specs/tid

     

       9
       9
       +
       

     

       10
       10
       +
         TID strings are always 13 characters long. All bits in the 64-bit number are

     

       11
       11
       +
         encoded, essentially meaning that the string is padded with "2" if necessary,

     

       12
       12
       +
         (the 0th character in the base32-sortable alphabet).

     

       13
       13
       +
         """

     

       14
       14
       +
         import Bitwise

     

       15
       15
       +
         alias Atex.Base32Sortable

     

       16
       16
       +
         use TypedStruct

     

       17
       17
       +
       

     

       18
       18
       +
         @re ~r/^[234567abcdefghij][234567abcdefghijklmnopqrstuvwxyz]{12}$/

     

       19
       19
       +
       

     

       20
       20
       +
         @typedoc """

     

       21
       21
       +
         A Unix timestamp representing when the TID was created.

     

       22
       22
       +
         """

     

       23
       23
       +
         @type timestamp() :: integer()

     

       24
       24
       +
       

     

       25
       25
       +
         @typedoc """

     

       26
       26
       +
         An integer to be used for the lower 10 bits of the TID.

     

       27
       27
       +
         """

     

       28
       28
       +
         @type clock_id() :: 0..1023

     

       29
       29
       +
       

     

       30
       30
       +
         typedstruct enforce: true do

     

       31
       31
       +
           field :timestamp, timestamp()

     

       32
       32
       +
           field :clock_id, clock_id()

     

       33
       33
       +
         end

     

       34
       34
       +
       

     

       35
       35
       +
         @doc """

     

       36
       36
       +
         Returns a TID for the current moment in time, along with a random clock ID.

     

       37
       37
       +
         """

     

       38
       38
       +
         @spec now() :: t()

     

       39
       39
       +
         def now,

     

       40
       40
       +
           do: %__MODULE__{

     

       41
       41
       +
             timestamp: DateTime.utc_now(:microsecond) |> DateTime.to_unix(:microsecond),

     

       42
       42
       +
             clock_id: gen_clock_id()

     

       43
       43
       +
           }

     

       44
       44
       +
       

     

       45
       45
       +
         @doc """

     

       46
       46
       +
         Create a new TID from a `DateTime` or an integer representing a Unix timestamp in microseconds.

     

       47
       47
       +
       

     

       48
       48
       +
         If `clock_id` isn't provided, a random one will be generated.

     

       49
       49
       +
         """

     

       50
       50
       +
         @spec new(DateTime.t() | integer(), integer() | nil) :: t()

     

       51
       51
       +
         def new(source, clock_id \\ nil)

     

       52
       52
       +
       

     

       53
       53
       +
         def new(%DateTime{} = datetime, clock_id),

     

       54
       54
       +
           do: %__MODULE__{

     

       55
       55
       +
             timestamp: DateTime.to_unix(datetime, :microsecond),

     

       56
       56
       +
             clock_id: clock_id || gen_clock_id()

     

       57
       57
       +
           }

     

       58
       58
       +
       

     

       59
       59
       +
         def new(unix, clock_id) when is_integer(unix),

     

       60
       60
       +
           do: %__MODULE__{timestamp: unix, clock_id: clock_id || gen_clock_id()}

     

       61
       61
       +
       

     

       62
       62
       +
         @doc """

     

       63
       63
       +
         Convert a TID struct to an instance of `DateTime`.

     

       64
       64
       +
         """

     

       65
       65
       +
         def to_datetime(%__MODULE__{} = tid), do: DateTime.from_unix(tid.timestamp, :microsecond)

     

       66
       66
       +
       

     

       67
       67
       +
         @doc """

     

       68
       68
       +
         Generate a random integer to be used as a `clock_id`.

     

       69
       69
       +
         """

     

       70
       70
       +
         @spec gen_clock_id() :: clock_id()

     

       71
       71
       +
         def gen_clock_id, do: :rand.uniform(1024) - 1

     

       72
       72
       +
       

     

       73
       73
       +
         @doc """

     

       74
       74
       +
         Decode a TID string into an `Atex.TID` struct, returning an error if it's invalid.

     

       75
       75
       +
       

     

       76
       76
       +
         ## Examples

     

       77
       77
       +
       

     

       78
       78
       +
         Syntactically valid TIDs:

     

       79
       79
       +
       

     

       80
       80
       +
             iex> Atex.TID.decode("3jzfcijpj2z2a")

     

       81
       81
       +
             {:ok, %Atex.TID{clock_id: 6, timestamp: 1688137381887007}}

     

       82
       82
       +
       

     

       83
       83
       +
             iex> Atex.TID.decode("7777777777777")

     

       84
       84
       +
             {:ok, %Atex.TID{clock_id: 165, timestamp: 5811096293381285}}

     

       85
       85
       +
       

     

       86
       86
       +
             iex> Atex.TID.decode("3zzzzzzzzzzzz")

     

       87
       87
       +
             {:ok, %Atex.TID{clock_id: 1023, timestamp: 2251799813685247}}

     

       88
       88
       +
       

     

       89
       89
       +
             iex> Atex.TID.decode("2222222222222")

     

       90
       90
       +
             {:ok, %Atex.TID{clock_id: 0, timestamp: 0}}

     

       91
       91
       +
       

     

       92
       92
       +
         Invalid TIDs:

     

       93
       93
       +
       

     

       94
       94
       +
             # not base32

     

       95
       95
       +
             iex> Atex.TID.decode("3jzfcijpj2z21")

     

       96
       96
       +
             :error

     

       97
       97
       +
             iex> Atex.TID.decode("0000000000000")

     

       98
       98
       +
             :error

     

       99
       99
       +
       

     

       100
       100
       +
             # case-sensitive

     

       101
       101
       +
             iex> Atex.TID.decode("3JZFCIJPJ2Z2A")

     

       102
       102
       +
             :error

     

       103
       103
       +
       

     

       104
       104
       +
             # too long/short

     

       105
       105
       +
             iex> Atex.TID.decode("3jzfcijpj2z2aa")

     

       106
       106
       +
             :error

     

       107
       107
       +
             iex> Atex.TID.decode("3jzfcijpj2z2")

     

       108
       108
       +
             :error

     

       109
       109
       +
             iex> Atex.TID.decode("222")

     

       110
       110
       +
             :error

     

       111
       111
       +
       

     

       112
       112
       +
             # legacy dash syntax *not* supported (TTTT-TTT-TTTT-CC)

     

       113
       113
       +
             iex> Atex.TID.decode("3jzf-cij-pj2z-2a")

     

       114
       114
       +
             :error

     

       115
       115
       +
       

     

       116
       116
       +
             # high bit can't be set

     

       117
       117
       +
             iex> Atex.TID.decode("zzzzzzzzzzzzz")

     

       118
       118
       +
             :error

     

       119
       119
       +
             iex> Atex.TID.decode("kjzfcijpj2z2a")

     

       120
       120
       +
             :error

     

       121
       121
       +
       

     

       122
       122
       +
         """

     

       123
       123
       +
         @spec decode(String.t()) :: {:ok, t()} | :error

     

       124
       124
       +
         def decode(<<timestamp::binary-size(11), clock_id::binary-size(2)>> = tid) do

     

       125
       125
       +
           if Regex.match?(@re, tid) do

     

       126
       126
       +
             timestamp = Base32Sortable.decode(timestamp)

     

       127
       127
       +
             clock_id = Base32Sortable.decode(clock_id)

     

       128
       128
       +
       

     

       129
       129
       +
             {:ok,

     

       130
       130
       +
              %__MODULE__{

     

       131
       131
       +
                timestamp: timestamp,

     

       132
       132
       +
                clock_id: clock_id

     

       133
       133
       +
              }}

     

       134
       134
       +
           else

     

       135
       135
       +
             :error

     

       136
       136
       +
           end

     

       137
       137
       +
         end

     

       138
       138
       +
       

     

       139
       139
       +
         def decode(_tid), do: :error

     

       140
       140
       +
       

     

       141
       141
       +
         @doc """

     

       142
       142
       +
         Encode a TID struct into a string.

     

       143
       143
       +
       

     

       144
       144
       +
         ## Examples

     

       145
       145
       +
       

     

       146
       146
       +
             iex> Atex.TID.encode(%Atex.TID{clock_id: 6, timestamp: 1688137381887007})

     

       147
       147
       +
             "3jzfcijpj2z2a"

     

       148
       148
       +
       

     

       149
       149
       +
             iex> Atex.TID.encode(%Atex.TID{clock_id: 165, timestamp: 5811096293381285})

     

       150
       150
       +
             "7777777777777"

     

       151
       151
       +
       

     

       152
       152
       +
             iex> Atex.TID.encode(%Atex.TID{clock_id: 1023, timestamp: 2251799813685247})

     

       153
       153
       +
             "3zzzzzzzzzzzz"

     

       154
       154
       +
       

     

       155
       155
       +
             iex> Atex.TID.encode(%Atex.TID{clock_id: 0, timestamp: 0})

     

       156
       156
       +
             "2222222222222"

     

       157
       157
       +
       

     

       158
       158
       +
         """

     

       159
       159
       +
         @spec encode(t()) :: String.t()

     

       160
       160
       +
         def encode(%__MODULE__{} = tid) do

     

       161
       161
       +
           timestamp = tid.timestamp |> Base32Sortable.encode() |> String.pad_leading(11, "2")

     

       162
       162
       +
           clock_id = (tid.clock_id &&& 1023) |> Base32Sortable.encode() |> String.pad_leading(2, "2")

     

       163
       163
       +
           timestamp <> clock_id

     

       164
       164
       +
         end

     

       165
       165
       +
       end

     

       166
       166
       +
       

     

       167
       167
       +
       defimpl String.Chars, for: Atex.TID do

     

       168
       168
       +
         def to_string(tid), do: Atex.TID.encode(tid)

     

       169
       169
       +
       end

test/tid_test.exs

···

       1
       1
       +
       defmodule TIDTest do

     

       2
       2
       +
         use ExUnit.Case, async: true

     

       3
       3
       +
         doctest Atex.TID

     

       4
       4
       +
       end