Get-FileCompressedMetaInfo.ps1

Add-PSSnapin Microsoft.SharePoint.Powershell

function ConvertFrom-HexStringToByteArray
{
    [CmdletBinding()]
    param
    (
        [parameter(Mandatory=$true)][string]$String
    )
    
    $bytes = New-Object byte[]($string.Length/2)
    
    for( $i = 0; $i -lt $String.Length; $i+=2 )
    {
        $bytes[$i/2] = [System.Convert]::ToByte( $string.Substring($i, 2), 16) 
    }

    return $bytes
}

function Decompress-String
{
    [CmdletBinding()]
    param
    (
        [parameter(Mandatory=$true)][string]$String
    )
    
    begin
    {
    }
    process
    {
        try
        {
            $String = $String -replace "^0x", ""

            $inputBytes = ConvertFrom-HexStringToByteArray -String $String

            $sourceStream      = New-Object System.IO.MemoryStream(,$inputBytes)
            $destinationStream = New-Object System.IO.MemoryStream

            # https://docs.microsoft.com/en-us/openspecs/sharepoint_protocols/ms-wssfo3/6ca01cb6-d75a-431e-8131-8c780ba46cd8
            
            # ID (2 bytes):  MUST be 0xA8A9.
            $sourceStream.ReadByte() | Out-Null # 0xA8
            $sourceStream.ReadByte() | Out-Null # 0xA9

            # Version (2 bytes):  MUST be 0x3031 (ASCII "01")
            $sourceStream.ReadByte() | Out-Null # 0x30
            $sourceStream.ReadByte() | Out-Null # 0x31

            # FileHeaderSize (4 bytes):  A 4-byte, unsigned integer specifying the size of the header. The value is 0x0C000000.
            $sourceStream.ReadByte() | Out-Null # 0x0C
            $sourceStream.ReadByte() | Out-Null # 0x00
            $sourceStream.ReadByte() | Out-Null # 0x00
            $sourceStream.ReadByte() | Out-Null # 0x00

            # OrigSize (4 bytes):  A 4-byte, unsigned integer specifying the size of the original content. It MUST be the size of the uncompressed stream before compression.
            $sourceStream.ReadByte() | Out-Null # 0xFC
            $sourceStream.ReadByte() | Out-Null # 0x00
            $sourceStream.ReadByte() | Out-Null # 0x00
            $sourceStream.ReadByte() | Out-Null # 0x00

            # ?
            $sourceStream.ReadByte() | Out-Null # 0x78
            $sourceStream.ReadByte() | Out-Null # 0x9C

            # zlib decompression
            $decompressionStream = New-Object System.IO.Compression.DeflateStream( $sourceStream, [System.IO.Compression.CompressionMode]::Decompress )
            $decompressionStream.CopyTo($destinationStream)

            # convert to string
            [System.text.encoding]::UTF8.GetString( $destinationStream.ToArray() )
        }
        finally
        {
            if( $destinationStream )
            {
                $destinationStream.Close()
                $destinationStream.Dispose()
            }

            if( $sourceStream )
            {
                $sourceStream.Close()
                $sourceStream.Dispose()
            }
        }
    }
    end
    {
    }
}

function Compress-String
{
    [CmdletBinding()]
    param
    (
        [parameter(Mandatory=$true)][string]$String
    )

        try
        {
            $String = $String -replace "\0", ""

            $String = ([Regex]::Replace($String, "(\r\n *<)", "<"))

            $inputBytes = [System.Text.Encoding]::UTF8.GetBytes($String)

            $originalFileStream = New-Object System.IO.MemoryStream(,$inputBytes  )
            $compressedStream   = New-Object System.IO.MemoryStream
            
            # zlib compression
            $compressor  = New-Object System.IO.Compression.DeflateStream( $compressedStream, [System.IO.Compression.CompressionMode]::Compress, $true )
            $originalFileStream.CopyTo( $compressor )
            $compressor.Close()

            $inputBytes2 = New-Object byte[] $($compressedStream.ToArray().Length + 14)

            $inputBytes2[0]  = 0xA8 # HEADER
            $inputBytes2[1]  = 0xA9 # HEADER

            $inputBytes2[2]  = 0x30 # VERSION
            $inputBytes2[3]  = 0x31 # VERSION

            $inputBytes2[4]  = 0x0C # HEADER SIZE
            $inputBytes2[5]  = 0x00 # HEADER SIZE
            $inputBytes2[6]  = 0x00 # HEADER SIZE
            $inputBytes2[7]  = 0x00 # HEADER SIZE

            $inputBytes2[8]  = $inputBytes.Length # ORIG SIZE
            $inputBytes2[9]  = 0x00 # ORIG SIZE
            $inputBytes2[10] = 0x00 # ORIG SIZE
            $inputBytes2[11] = 0x00 # ORIG SIZE

            $inputBytes2[13] = 0x78 # ?
            $inputBytes2[14] = 0x96 # ?

            $compressedStream.ToArray().CopyTo( $inputBytes2, 14 )

            ConvertFrom-ArrayToHexString -Array $inputBytes2
        }
        finally
        {
            if( $deflator )
            {
                $deflator.Close()
                $deflator.Dispose()
            }

            if( $originalFileStream )
            {
                $originalFileStream.Close()
                $originalFileStream.Dispose()
            }

            if( $compressedStream )
            {
                $compressedStream.Close()
                $compressedStream.Dispose()
            }
        }
}

function Get-DataTable
{
    [cmdletbinding()]
    param
    (
        [Parameter(Mandatory=$true,ParameterSetName="Individual")]
        [string]$DatabaseName,

        [Parameter(Mandatory=$true,ParameterSetName="Individual")]
        [string]$DatabaseServer,

        [Parameter(Mandatory=$true,ParameterSetName="ConnectionString")]
        [string]$ConnectionString,


        [Parameter(Mandatory=$true,ParameterSetName="Individual")]
        [Parameter(Mandatory=$true,ParameterSetName="ConnectionString")]
        [string]$Query,

        [Parameter(Mandatory=$false,ParameterSetName="Individual")]
        [Parameter(Mandatory=$false,ParameterSetName="ConnectionString")]
        [int]$CommandTimeout=30, # The default is 30 seconds

        [Parameter(Mandatory=$false,ParameterSetName="Individual")]
        [Parameter(Mandatory=$false,ParameterSetName="ConnectionString")]
        [HashTable]$Parameters = @{}
    )

    begin
    {
        if( $PSCmdlet.ParameterSetName -eq "Individual" )
        {
            $ConnectionString = "Data Source=$DatabaseServer;Initial Catalog=$DatabaseName;Integrated Security=True;Enlist=False;Connect Timeout=5"
        }
    }
    process
    {
        try
        {
            $dataSet     = New-Object System.Data.DataSet     
            $dataAdapter = New-Object System.Data.SqlClient.SqlDataAdapter( $Query, $ConnectionString )
            
            foreach( $Parameter in $Parameters.GetEnumerator() )
            {
                $param = $dataAdapter.SelectCommand.Parameters.AddWithValue( "@$($Parameter.Key)", $Parameter.Value )
            }

            Write-Debug -Message "$(Get-Date) - Executing Query: $Query"

            $dataAdapter.Fill($dataSet) | Out-Null
            return $dataSet.Tables[0]
        }
        catch
        {
            throw $_.Exception
        }
        finally
        {
            if($dataSet)
            {
                $dataSet.Dispose()
            }

            if($dataAdapter)
            {
                $dataAdapter.Dispose()
            }
        }
    }
    end
    {
    }
}

function Invoke-NonQuery
{
    [cmdletbinding()]
    param
    (
        [Parameter(Mandatory=$true,ParameterSetName="Individual")]
        [string]$DatabaseName,

        [Parameter(Mandatory=$true,ParameterSetName="Individual")]
        [string]$DatabaseServer,

        [Parameter(Mandatory=$true,ParameterSetName="ConnectionString")]
        [string]$ConnectionString,


        [Parameter(Mandatory=$true,ParameterSetName="Individual")]
        [Parameter(Mandatory=$true,ParameterSetName="ConnectionString")]
        [string]$Query,

        [Parameter(Mandatory=$false,ParameterSetName="Individual")]
        [Parameter(Mandatory=$false,ParameterSetName="ConnectionString")]
        [int]$CommandTimeout=30, # The default is 30 seconds

        [Parameter(Mandatory=$false,ParameterSetName="Individual")]
        [Parameter(Mandatory=$false,ParameterSetName="ConnectionString")]
        [HashTable]$Parameters = @{}
    )

    begin
    {
        if( $PSCmdlet.ParameterSetName -eq "Individual" )
        {
            $ConnectionString = "Data Source=$DatabaseServer;Initial Catalog=$DatabaseName;Integrated Security=True;Enlist=False;Connect Timeout=5"
        }
    }
    process
    {
        try
        {
            $connection = New-Object System.Data.SqlClient.SqlConnection($ConnectionString)
            $connection.Open()

            $command = New-Object system.Data.SqlClient.SqlCommand($Query, $connection)     
            $command.CommandTimeout = $CommandTimeout

            foreach( $Parameter in $Parameters.GetEnumerator() )
            {
                $param = $command.Parameters.Add( "@$($Parameter.Key)", $Parameter.Value )
            }

            $command.ExecuteNonQuery() | Out-Null
        }
        catch
        {
            throw $_.Exception
        }
        finally
        {
            if($connection)
            {
                [System.Data.SqlClient.SqlConnection]::ClearAllPools()
                $connection.Close()
                $connection.Dispose()
            }
        }
    }
    end
    {
    }
}

function Get-FileMetaInfo
{
    [CmdletBinding()]
    param
    (
        [parameter(Mandatory=$true)][string]$FileUrl
    )

    begin
    {
    }
    process
    {
        try
        {
            $site = New-Object Microsoft.SharePoint.SPSite($FileUrl)
        
            $web = $site.Openweb()

            $file = $web.GetFile($FileUrl)

            if( $file.Exists )
            {
                $row = Get-DataTable `
                        -DatabaseName   $site.ContentDatabase.Name `
                        -DatabaseServer $site.ContentDatabase.Server `
                        -Query          "SELECT Id, CONVERT(nvarchar(max), MetaInfo, 1) AS 'CompressedMetaInfo' FROM [AllDocs] WHERE SiteId = @SiteId AND Id = @UniqueId" `
                        -Parameters     @{ SiteId = $site.id; UniqueId = $file.UniqueId }

                [PSCustomObject] @{
                    Site                 = $site.Url
                    Web                  = $web.Url
                    FileUrl              = $site.MakeFullUrl($File.Url)
                    Author               = $file.Author.LoginName
                    Editor               = $file.ModifiedBy.LoginName
                    vti_author           = $file.Properties["vti_author"]
                    vti_modifiedby       = $file.Properties["vti_modifiedby"]
                    CompressedMetaInfo   = $row.CompressedMetaInfo
                    DecompressedMetaInfo = (Decompress-String -String $row.CompressedMetaInfo)
                }
            }
            else
            {
                throw New-Object System.IO.FileNotFoundException("The file at '$fileUrl' was not found.", $FileUrl)
            }
        }
        finally
        {
            if( $web )
            {
                $web.Dispose()
            }

            if( $site )
            {
                $site.Dispose()
            }
        }
    }
    end
    {
    }
}

function Set-FileMetaInfo
{
    [CmdletBinding()]
    param
    (
        [parameter(Mandatory=$true)][string]$FileUrl,
        [parameter(Mandatory=$true)][string]$CompressedMetaInfo
    )

    begin
    {
    }
    process
    {
        try
        {
            $site = New-Object Microsoft.SharePoint.SPSite($FileUrl)
        
            $web = $site.Openweb()

            $file = $web.GetFile($FileUrl)

            if( $file.Exists )
            {
                Invoke-NonQuery `
                    -DatabaseName   $site.ContentDatabase.Name `
                    -DatabaseServer $site.ContentDatabase.Server `
                    -Query          "UPDATE Alldocs SET MetaInfo = CONVERT(varbinary(max), '$CompressedMetaInfo', 1) WHERE SiteId = @SiteId AND Id = @UniqueId" `
                    -Parameters     @{ SiteId = $site.Id; UniqueId = $file.UniqueId } 
            }
            else
            {
                throw New-Object System.IO.FileNotFoundException("The file at '$fileUrl' was not found.", $FileUrl)
            }
        }
        finally
        {
            if( $web )
            {
                $web.Dispose()
            }

            if( $site )
            {
                $site.Dispose()
            }
        }
    }
    end
    {
    }
}


Get-FileMetaInfo -FileUrl "https://sharepoint.2016.contoso.com/sites/teamsite/Shared%20Documents/Folder%201/Folder%202/Folder%203/Folder%204/ftype.txt"