eunzip's Introduction

eunzip

Simple unzip library with Zip64 support

Overview

Eunzip allows opening, verifying and decompressing of Zip files > 2GB and Zip64 archives, which are not supported by the standard Erlang zip module.

Supports STORE and DEFLATE compression methods.

Basic usage

% Open a Zip file and read its structure
{ok, UnzipState} = eunzip:open("foo/bar.zip")

% Get all file and directory entries from a Zip archive
{ok, Entries} = eunzip:entries(UnzipState)

% Get specific file or directory entry from a Zip archive
{ok, Entry} = eunzip:entry(UnzipState, "README.md")

% Verify an archived file checksum
ok = eunzip:verify(UnzipState, "README.md")

% Decompress file from a Zip archive into as a target filename
ok = eunzip:decompress(UnzipState, "README.md", "unpacked/README.md")

% Close a previously opened Zip file
ok = eunzip:close(UnzipState)

Note: closing a previously opened Zip file is mandatory.

File streaming

Archived file streaming is a technique allowing to process decompressed file by chunks and perform arbitrary actions.

The example below calculates MD5 checksum on-the-fly.

stream_md5(ZipFile, ArchivedFile) ->
    {ok, UnzipState} = eunzip:open("foo/bar.zip"),
    {ok, StreamState} = eunzip:stream_init(UnzipState, "hugefile.mkv"),
    {ok, Md5} = stream_fun(StreamState, crypto:hash_init(md5)),
    ok = eunzip:close(UnzipState),
    Md5.

stream_fun(StreamState, Md5State) ->
    % Read 1 MB chunks of archived data per call
    case eunzip:stream_read_chunk(1024 * 1024, StreamState) of
        {ok, Data} ->
            Md5State1 = crypto:hash_update(Md5State, Data),
            {ok, crypto:hash_final(Md5State1)};
        {more, Data, StreamState1} ->
            stream_fun(StreamState1, crypto:hash_update(Md5State, Data));
        {error, Reason} ->
            {error, Reason}
    end.

Note: stream is closed automatically if all of the file data is read or an error occured.

If you don't want to read until the end of the file, you should call eunzip:stream_end/1 function:

ok = eunzip:stream_end(StreamState)

Credits

Eunzip is based on the Unzip Elixir library by Akash Hiremath.

eunzip's People

Contributors

Stargazers

Watchers

eunzip's Issues

Some dialyzer problems around opaque terms.

Hi,

eunzip is a nice library.

While testing it I've found the following dialyzer errors:

Erlang 24.3.4.13

/home/joaohf/opensource/eunzip/src/eunzip.erl:78:25: The pattern
          {'ok', CD} can never match the type
          {'error', atom()}

/home/joaohf/opensource/eunzip/src/eunzip_central_dir.erl:33:9: The pattern
          {'error', Reason} can never match the type
          {'ok',
           #file_buffer{file :: {'file_descriptor', atom(), _},
                        size :: non_neg_integer(),
                        limit :: non_neg_integer(),
                        buffer :: <<>>,
                        buffer_size :: non_neg_integer(),
                        buffer_position :: non_neg_integer(),
                        direction :: 'backward'}}

/home/joaohf/opensource/eunzip/src/eunzip_central_dir.erl:154:25: The variable Acc1 can never match since previous clauses completely covered the type
          {'error', 'overlapped_zip_entries'}

/home/joaohf/opensource/eunzip/src/eunzip_central_dir.erl:167:34: The call eunzip_range:is_overlap
         (RangeTree :: gb_trees:tree(_, _),
          HeaderOffset :: non_neg_integer(),
          CompressedSize :: non_neg_integer()) contains an opaque term as 1st argument when terms of different types are expected in these positions

/home/joaohf/opensource/eunzip/src/eunzip_range.erl:28:2: Invalid type specification for function eunzip_range:is_overlap/3. The success typing is
          ({_, 'nil' | {_, _, _, _}}, _, _) -> boolean()

Erlang 26.0.2

/home/joaohf/opensource/eunzip/src/eunzip.erl:78:25: The pattern
          {'ok', CD} can never match the type
          {'error', atom()}

/home/joaohf/opensource/eunzip/src/eunzip_central_dir.erl:33:9: The pattern
          {'error', Reason} can never match the type
          {'ok',
           #file_buffer{file :: {'file_descriptor', atom(), _},
                        size :: non_neg_integer(),
                        limit :: non_neg_integer(),
                        buffer :: <<>>,
                        buffer_size :: non_neg_integer(),
                        buffer_position :: non_neg_integer(),
                        direction :: 'backward'}}

/home/joaohf/opensource/eunzip/src/eunzip_central_dir.erl:154:25: The variable Acc1 can never match since previous clauses completely covered the type
          {'error', 'overlapped_zip_entries'}

/home/joaohf/opensource/eunzip/src/eunzip_central_dir.erl:167:34: The call eunzip_range:is_overlap
         (RangeTree :: gb_trees:tree(_, _),
          HeaderOffset :: non_neg_integer(),
          CompressedSize :: non_neg_integer()) contains an opaque term as 1st argument when terms of different types are expected in these positions

/home/joaohf/opensource/eunzip/src/eunzip_range.erl:28:2: Invalid type specification for function eunzip_range:is_overlap/3.
 The success typing is eunzip_range:is_overlap
          ({_, 'nil' | {_, _, _, _}}, _, _) -> boolean()
 But the spec is eunzip_range:is_overlap
          (Tree, Offset, Length) -> Result
             when
                 Tree :: gb_trees:tree(),
                 Offset :: non_neg_integer(),
                 Length :: non_neg_integer(),
                 Result :: boolean()
 They do not overlap in the 1st argument

Recommend Projects