I was cleaning up and organizing bookmarks in Google Chrome today and decided to find out where they were stored on my computer. I found the Bookmarks file in a local app data folder. Opening it up in Notepad I was pleasantly surprised to discover it is in JSON. Excellent! This gives me an opportunity to try out some of the new web cmdlets in PowerShell v3 and build a little tool to help find broken links.<\/p>\n
The first step is to convert the file from JSON into PowerShell objects.<\/p>\n
$File = \"$env:localappdata\\Google\\Chrome\\User Data\\Default\\Bookmarks\"\r\n$data = Get-content $file | out-string | ConvertFrom-Json\r\n<\/pre>\nTo convert from JSON, the input needs to be one long string. Get-Content writes an array of strings so by piping to Out-String first, ConvertFrom-JSON is happy. Here's what I end up with.<\/p>\n
$data\r\n\r\nchecksum roots version\r\n-------- ----- -------\r\n03a5a00f42bb4860f6f8dd4d543e34af @{bookmark_bar=; other=; synced=} 1\r\n<\/pre>\nThe roots property is where things are stored.<\/p>\n
$data.roots | format-list\r\n\r\nbookmark_bar : @{children=System.Object[]; date_added=12989558548133917; date_modified=12998163702118866; id=1;\r\n name=Bookmarks bar; type=folder}\r\nother : @{children=System.Object[]; date_added=12989558548133917; date_modified=12998151911083334; id=2;\r\n name=Other bookmarks; type=folder}\r\nsynced : @{children=System.Object[]; date_added=12989558548133917; date_modified=0; id=3; name=Mobile bookmarks;\r\n type=folder}\r\n<\/pre>\nAs far as I know these are hard-coded properties. Each property can have a nested property called children which will be a collection of bookmarks and subfolders.<\/p>\n
$data.roots.bookmark_bar\r\n\r\n\r\nchildren : {@{date_added=12993566428951635; id=67; name=PowerShell.org \u2022 View unanswered posts; type=url;\r\n url=http:\/\/powershell.org\/discuss\/search.php?search_id=unanswered}, @{date_added=12989558569500214;\r\n id=5; name=HostGator.com Control Panel; type=url;\r\n url=http:\/\/gator1172.hostgator.com:2082\/frontend\/x3\/index.php?post_login=91229198020615},\r\n @{date_added=12989558569506214; id=7; name=Vet Followers; type=url;\r\n url=https:\/\/www.socialoomph.com\/vetfollowers}, @{date_added=12989558569508214; id=8; name=Google+;\r\n type=url;\r\n url=https:\/\/plus.google.com\/up\/start\/?continue=https:\/\/plus.google.com\/&type=st&gpcaz=3cba226a}...}\r\ndate_added : 12989558548133917\r\ndate_modified : 12998163702118866\r\nid : 1\r\nname : Bookmarks bar\r\ntype : folder\r\n<\/pre>\nHere's what an child item looks like:<\/p>\n
date_added : 12998151910630334
\nid : 76
\nname : Facebook
\ntype : url
\nurl : http:\/\/www.facebook.com\/<\/p>\nAwesome. All I need to do is get the url.<\/p>\n
$data.roots.bookmark_bar.children | select Name,url\r\n\r\nname url\r\n---- ---\r\nPowerShell.org \u2022 View unanswered posts http:\/\/powershell.org\/discuss\/search.php?search_id=unans...\r\nHostGator.com Control Panel http:\/\/gator1172.hostgator.com:2082\/frontend\/x3\/index.ph...\r\nVet Followers https:\/\/www.socialoomph.com\/vetfollowers\r\nGoogle+ https:\/\/plus.google.com\/up\/start\/?continue=https:\/\/plus....\r\nBottlenose http:\/\/bottlenose.com\/home#streams\/everything\r\nPower Tweet javascript:(function(){url=\"http:\/\/www.twylah.com\/bookma...\r\nFacebook http:\/\/www.facebook.com\/\r\nBlog Dashboard http:\/\/jdhitsolutions.com\/blog\/wp-admin\/index.php\r\n<\/pre>\nTo validate if the URL is good, I can use Invoke-Webrequest.<\/p>\n
invoke-webrequest -Uri http:\/\/www.facebook.com -UseBasicParsing | Select StatusCode\r\n StatusCode\r\n ----------\r\n 200\r\n<\/pre>\nAll I want is the status code so I'm using basic parsing to speed things up. Now that I have the basics, I can turn this into a script.<\/p>\n
#requires -version 3.0\r\n\r\n#comment based help is here\r\n\r\n[cmdletbinding()]\r\n\r\nParam (\r\n[Parameter(Position=0)]\r\n[ValidateScript({Test-Path $_})]\r\n[string]$File = \"$env:localappdata\\Google\\Chrome\\User Data\\Default\\Bookmarks\",\r\n[switch]$Validate\r\n)\r\n\r\nWrite-Verbose -Message \"Starting $($MyInvocation.Mycommand)\"\r\n\r\n#A nested function to enumerate bookmark folders\r\nFunction Get-BookmarkFolder {\r\n[cmdletbinding()]\r\nParam(\r\n[Parameter(Position=0,ValueFromPipeline=$True)]\r\n$Node\r\n)\r\n\r\nProcess {\r\n \r\n foreach ($child in $node.children) {\r\n #get parent folder name\r\n $parent = $node.Name\r\n if ($child.type -eq 'Folder') {\r\n Write-Verbose \"Processing $($child.Name)\"\r\n Get-BookmarkFolder $child\r\n }\r\n else {\r\n $hash= [ordered]@{\r\n Folder = $parent\r\n Name = $child.name\r\n URL = $child.url\r\n Added = [datetime]::FromFileTime(([double]$child.Date_Added)*10)\r\n Valid = $Null\r\n Status = $Null\r\n }\r\n If ($Validate) {\r\n Write-Verbose \"Validating $($child.url)\"\r\n if ($child.url -match \"^http\") {\r\n #only test if url starts with http or https\r\n Try {\r\n $r = Invoke-WebRequest -Uri $child.url -DisableKeepAlive -UseBasicParsing\r\n if ($r.statuscode -eq 200) {\r\n $hash.Valid = $True\r\n } #if statuscode\r\n else {\r\n $hash.valid = $False\r\n }\r\n $hash.status = $r.statuscode\r\n Remove-Variable -Name r -Force\r\n }\r\n Catch {\r\n Write-Warning \"Could not validate $($child.url)\"\r\n $hash.valid = $False\r\n $hash.status = $Null\r\n }\r\n \r\n } #if url\r\n \r\n } #if validate\r\n #write custom object\r\n New-Object -TypeName PSobject -Property $hash\r\n } #else url\r\n } #foreach\r\n } #process\r\n} #end function\r\n\r\n#convert Google Chrome Bookmark filefrom JSON\r\n$data = Get-Content $file | Out-String | ConvertFrom-Json\r\n\r\n#these should be the top level \"folders\"\r\n$data.roots.bookmark_bar | Get-BookmarkFolder \r\n$data.roots.other | Get-BookmarkFolder \r\n$data.roots.synced | Get-BookmarkFolder\r\n\r\nWrite-Verbose -Message \"Ending $($MyInvocation.Mycommand)\"\r\n<\/pre>\nThis script converts the bookmarks file from JSON and creates custom objects for each bookmark using the nested Get-BookMarkFolder function. This processes each child if it is a url. If it is a folder then it passes the folder name recursively to Get-BookmarkFolder. Because validation might be time consuming, I made it optional with -Validate. I also converted the date_added property into a user-friendly datetime format. The value in the file is a file time, i.e. number of ticks since 1\/1\/1601. Although the actual value needs to be multiplied by 10 to get the correct date time.<\/p>\n
When I run the script, without validation I get an object like this for each bookmark.<\/p>\n
Folder : Misc
\nName : [Release][Alpha0.6] CyanogenMod 9 Touchpad - RootzWiki
\nURL : http:\/\/rootzwiki.com\/topic\/15509-releasealpha06-cyanogenmod-9-touchpad\/
\nAdded : 8\/15\/2012 10:42:49 PM
\nValid :
\nStatus :<\/p>\nIf I validate it will look like this:<\/p>\n
Folder : Misc
\nName : [Release][Alpha0.6] CyanogenMod 9 Touchpad - RootzWiki
\nURL : http:\/\/rootzwiki.com\/topic\/15509-releasealpha06-cyanogenmod-9-touchpad\/
\nAdded : 8\/15\/2012 10:42:49 PM
\nValid : True
\nstatus : 200<\/p>\nBookmarks that fail will show as False. I haven't gotten around to figuring out how to rewrite the bookmarks file, but I would probably end up using ConvertTo-JSON.<\/p>\n
The web cmdlets in PowerShell v3 can be a lot of fun to work with. But one word of caution: do NOT try to turn the script in the ISE using -validate. It is possible that Invoke-WebRequest will begin consuming all memory on your computer, unless you make sure the ISE process is completely killed.<\/p>\n