更多我的博客请关注：https://kekbin.com/
在IPFS实践之初体验中以命令行的方式演示了如何使用IPFS的基础命令。然而，如果基于IPFS做二次开发应用程序，就需要对这些功能进行封装。本文介绍两种调用IPFS功能的方式。
Cmd
以C#为例，采用启动cmd进程的方式调用命令行,命令行执行程序如下：
public const int WAIT_FOR_EXIT = 2 * 60 * 000;
public const int CRASH_EXIT_CODE = 0xFFFF;
public static int RunExe(string exeFileName, string args, out string output)
{
output = "";
int iCode = CRASH_EXIT_CODE;
if (exeFileName == null)
{
return CRASH_EXIT_CODE;
}
ProcessStartInfo procInfo = new ProcessStartInfo
{
FileName = exeFileName,
Arguments = args,
RedirectStandardOutput = true,
CreateNoWindow = true,
UseShellExecute = false,
};
Process exeProc = null;
StreamReader reader = null;
try
{
exeProc = Process.Start(procInfo);
if (exeProc == null)
{
return CRASH_EXIT_CODE;
}
if (exeProc.StandardOutput != null && exeProc.StandardOutput.BaseStream != null)
{
output = exeProc.StandardOutput.ReadToEnd();
}
if (exeProc != null)
{
if (exeProc.WaitForExit(WAIT_FOR_EXIT))
{
iCode = exeProc.ExitCode;
}
int cnt = 0;
while (!exeProc.HasExited && cnt++ < 15)
{
Thread.Sleep(5);
}
exeProc.Close();
}
}
catch (Exception e)
{
iCode = CRASH_EXIT_CODE;
}
finally
{
DisposeResource(reader);
}
return iCode;
}
有了以上的封装，那么执行ipfs add的调用代码可以是
string args = string.Format("add {0}", filepath)
ScriptTool.RunExe("ipfs", args, out output);
这样的方式，有几种弊端：
- 由于每执行一条命令都需要启动一个进程，而这些进程是互斥的，在IPFS repo下，有一个repo_lock,导致同时命令多条执行的失败率很高。
- 由于命令行执行时，执行完成的输出都是字符串，处理起来比较麻烦。
Http Api
之前提过，官方提供的go-ipfs实现的demo中提供了http api的功能，也就是说所有的IPFS功能都可以通过http请求的方式来实现，http接口详见http API docs
以http api的方式执行，可以完美解决cmd的问题：
- 它无需另起进程，可以使用多线程，同时执行多条指令
- api的返回是一个json string，方便解析我们所需要的结果参数。
还是以ipfs add为例，接口说明如下：
/api/v0/add
Add a file or directory to ipfs.
Arguments
arg [file]: The path to a file to be added to ipfs. Required: yes.
recursive [bool]: Add directory paths recursively. Default: “false”. Required: no.
quiet [bool]: Write minimal output. Required: no.
quieter [bool]: Write only final hash. Required: no.
silent [bool]: Write no output. Required: no.
progress [bool]: Stream progress data. Required: no.
trickle [bool]: Use trickle-dag format for dag generation. Required: no.
only-hash [bool]: Only chunk and hash - do not write to disk. Required: no.
wrap-with-directory [bool]: Wrap files with a directory object. Required: no.
hidden [bool]: Include files that are hidden. Only takes effect on recursive add. Required: no.
chunker [string]: Chunking algorithm to use. Required: no.
pin [bool]: Pin this object when adding. Default: “true”. Required: no.
raw-leaves [bool]: Use raw blocks for leaf nodes. (experimental). Required: no.
nocopy [bool]: Add the file using filestore. (experimental). Required: no.
fscache [bool]: Check the filestore for pre-existing blocks. (experimental). Required: no.
cid-version [int]: Cid version. Non-zero value will change default of ‘raw-leaves’ to true. (experimental). Default: “0”. Required: no.
hash [string]: Hash function to use. Will set Cid version to 1 if used. (experimental). Default: “sha2-256”. Required: no.
Request Body
Argument “path” is of file type. This endpoint expects a file in the body of the request as ‘multipart/form-data’.
Response
On success, the call to this endpoint will return with 200 and the following body:
{
"Name": "[string]"
"Hash": "[string]"
"Bytes": "[int64]"
"Size": "[string]"
}
所以add一个文件,需要用post的方式，通过mutipart/form-data格式提交文件内容
这个请求的url：
http://localhost:5001/api/v0/add?recursive=false
其中上传文件夹的时候，recursive参数需要设置为true。
Mutipart上传代码如下
public static string HttpMultiPartPost(string url, int timeOut, string path, bool isDir)
{
string responseContent;
var webRequest = (HttpWebRequest)WebRequest.Create(url);
// 边界符
var boundary = "---------------" + DateTime.Now.Ticks.ToString("x");
// 边界符
var beginBoundary = Encoding.ASCII.GetBytes("
" + "--" + boundary + "
");
// 最后的结束符
var endBoundary = Encoding.ASCII.GetBytes("
" + "--" + boundary + "--
");
// 设置属性
webRequest.Method = "POST";
webRequest.Timeout = timeOut;
webRequest.ContentType = "multipart/form-data; boundary=" + boundary;
var requestStream = webRequest.GetRequestStream();
if (isDir)
{
//发送分割符
requestStream.Write(beginBoundary, 0, beginBoundary.Length);
//发送文件夹头
const string folderHeaderFormat =
"Content-Disposition: file; filename="{0}"
" +
"Content-Type: application/x-directory
";
var folderHeader = string.Format(folderHeaderFormat, GetDirectoryName(path));
var headerbytes = Encoding.UTF8.GetBytes(folderHeader);
requestStream.Write(headerbytes, 0, headerbytes.Length);
DirectoryInfo directory = new DirectoryInfo(path);
FileInfo[] fileInfo = directory.GetFiles();
foreach (FileInfo item in fileInfo)
{
PostFileItem(requestStream, beginBoundary, item, GetDirectoryName(path));
}
}
else
{
PostFileItem(requestStream, beginBoundary, new FileInfo(path), null);
}
// 写入最后的结束边界符
requestStream.Write(endBoundary, 0, endBoundary.Length);
requestStream.Close();
var httpWebResponse = (HttpWebResponse)webRequest.GetResponse();
using (var httpStreamReader = new StreamReader(httpWebResponse.GetResponseStream(),Encoding.GetEncoding("utf-8")))
{
responseContent = httpStreamReader.ReadToEnd();
}
httpWebResponse.Close();
webRequest.Abort();
return responseContent;
}
private static void PostFileItem(Stream stream, byte[] boundary, FileInfo fileInfo, string dirName)
{
string name = fileInfo.Name;
if (!string.IsNullOrEmpty(dirName))
{
name = dirName + "/" + name;
}
//发送分割符
stream.Write(boundary, 0, boundary.Length);
//发送文件头
const string headerFormat =
"Abspath: {0}
" +
"Content-Disposition: file; filename="{1}"
" +
"Content-Type: application/octet-stream
";
var header = string.Format(headerFormat, fileInfo.FullName, name);
var headerbytes = Encoding.UTF8.GetBytes(header);
stream.Write(headerbytes, 0, headerbytes.Length);
//发送文件流
var fileStream = new FileStream(fileInfo.FullName, FileMode.Open, FileAccess.Read);
var buffer = new byte[1024];
int bytesRead; // =0
while ((bytesRead = fileStream.Read(buffer, 0, buffer.Length)) != 0)
{
stream.Write(buffer, 0, bytesRead);
}
fileStream.Close();
}
add Api调用代码如下：
public static string Add(string path, bool isDir)
{
if (string.IsNullOrEmpty(path))
{
return null;
}
string format = "";
if (isDir)
{
format = "{0}?chunker=size-262144&recursive=true";
}
else
{
//如果需要保存名字，需要wrap一个dir
//format = "{0}?chunker=size-262144&recursive=false&wrap-with-directory=true";
format = "{0}?chunker=size-262144&recursive=false";
}
string url = string.Format(format, "http://localhost:5001/api/v0/add");
string output = HttpMultiPartPost(url, 500000, path, isDir);
return output;
}
大部分的api都是以get的方式请求，相比add简单许多，这里提供一个C#的httpget方法供参考.
public static string HttpGet(string url, int timeOut)
{
string responseContent;
var webRequest = (HttpWebRequest)WebRequest.Create(url);
webRequest.Method = "GET";
webRequest.Timeout = timeOut;
var httpWebResponse = (HttpWebResponse)webRequest.GetResponse();
using (var httpStreamReader = new StreamReader(httpWebResponse.GetResponseStream(), Encoding.GetEncoding("utf-8")))
{
responseContent = httpStreamReader.ReadToEnd();
}
httpWebResponse.Close();
webRequest.Abort();
return responseContent;
}
如name publish这样的操作就可以这样调用
public static string NamePublish(string hash, int hours)
{
//lifetime失效期
string format = "{0}?arg={1}&lifetime={2}h";
string url = string.Format(format, @"http://localhost:5001/api/v0/name/publish", hash, hours);
return HttpTool.HttpGet(url, 100000);
}
封装了这些操作，我们就可以将IPFS用起来，做一些我们所需要的application。