feat!(c++): Refactor code structure to consolidate include and src into a single src directory

[acezen]

Reason for this PR

Our current project structure separates header files and source files into include and src directories, respectively. This approach, while beneficial in some contexts, adds complexity and overhead for smaller projects, and not fits to protobuf generated code. To simplify the structure and enhance maintainability, we need to consolidate all code into a single src directory.

What changes are included in this PR?

Here are some notable changes:

• Consolidate the header and source code into single src directory.
• Add subdirectory to classify the arrow and high-level related code.
• Add api headers to let users to easily to include the function of c++ library.

Are these changes tested?

Yes

Are there any user-facing changes?

BREAKING CHANGE:

• the header path and header name may change, like:
  • graphar/arrow_chunk_reader.h to graphar/arrow/chunk_reader.h
  • graphar/arrow_chunk_writer.h to graphar/arrow/chunk_writer.h
• The copy way writer function has been moved from arrow_chunk_writer.h to chunk_info_writer.h

[acezen]

Hi, @lixueclaire, I create this PR to consolidate all code into a single src directory and brings some BREAKING CHANGE, may need your opinion about the changes, especially the naming of the directory and file:

• I use high-level for our high level reader and writer directory name
• Since we use arrow folder to separate the arrow reader and writer, so I change the arrow_chunk_reader and arrow_chunk_writer to chunk_reader and chunk_writer.
• And also the api 's header names.

[lixueclaire]

Hi, @lixueclaire, I create this PR to consolidate all code into a single src directory and brings some BREAKING CHANGE, may need your opinion about the changes, especially the naming of the directory and file:

• I use high-level for our high level reader and writer directory name
• Since we use arrow folder to separate the arrow reader and writer, so I change the arrow_chunk_reader and arrow_chunk_writer to chunk_reader and chunk_writer.
• And also the api 's header names.

Hi, I think the APIs maybe a little confusing now. Originally, we have chunk_info_reader and arrow_chunk_reader and high-level graph collections. If we use basic_reader, it is not clear what is read. I suggest naming info_reader/meta_reader,arrow_reader/chunk_reader and high_level_reader/graph_reader.

[acezen]

Hi, @lixueclaire, I create this PR to consolidate all code into a single src directory and brings some BREAKING CHANGE, may need your opinion about the changes, especially the naming of the directory and file:

• I use high-level for our high level reader and writer directory name
• Since we use arrow folder to separate the arrow reader and writer, so I change the arrow_chunk_reader and arrow_chunk_writer to chunk_reader and chunk_writer.
• And also the api 's header names.

Hi, I think the APIs maybe a little confusing now. Originally, we have chunk_info_reader and arrow_chunk_reader and high-level graph collections. If we use basic_reader, it is not clear what is read. I suggest naming info_reader/meta_reader,arrow_reader/chunk_reader and high_level_reader/graph_reader.

You mean we should split reader and writer to info, arrow and high_leveland also split the same level reader/writer to different directory?

[lixueclaire]

Hi, @lixueclaire, I create this PR to consolidate all code into a single src directory and brings some BREAKING CHANGE, may need your opinion about the changes, especially the naming of the directory and file:

• I use high-level for our high level reader and writer directory name
• Since we use arrow folder to separate the arrow reader and writer, so I change the arrow_chunk_reader and arrow_chunk_writer to chunk_reader and chunk_writer.
• And also the api 's header names.

Hi, I think the APIs maybe a little confusing now. Originally, we have chunk_info_reader and arrow_chunk_reader and high-level graph collections. If we use basic_reader, it is not clear what is read. I suggest naming info_reader/meta_reader,arrow_reader/chunk_reader and high_level_reader/graph_reader.

You mean we should split reader and writer to info, arrow and high_leveland also split the same level reader/writer to different directory?

I'm talking about the file names under the api directory.

[acezen]

Hi, @lixueclaire, I create this PR to consolidate all code into a single src directory and brings some BREAKING CHANGE, may need your opinion about the changes, especially the naming of the directory and file:

• I use high-level for our high level reader and writer directory name
• Since we use arrow folder to separate the arrow reader and writer, so I change the arrow_chunk_reader and arrow_chunk_writer to chunk_reader and chunk_writer.
• And also the api 's header names.

Hi, I think the APIs maybe a little confusing now. Originally, we have chunk_info_reader and arrow_chunk_reader and high-level graph collections. If we use basic_reader, it is not clear what is read. I suggest naming info_reader/meta_reader,arrow_reader/chunk_reader and high_level_reader/graph_reader.

You mean we should split reader and writer to info, arrow and high_leveland also split the same level reader/writer to different directory?

I'm talking about the file names under the api directory.

I have changed the naming under api, could you help reviewing again?

[lixueclaire]

LGTM now. And it's essential to update our documentation accordingly.