switch to python and combineHEADmain

6 files changed, 237 insertions, 156 deletions

diff --git a/README.md b/README.md
index b52b484..1251159 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,6 @@
 # Teams to rclone
 
-powershell script to collect all the drive_ids in sharepoint associated to a team and create a union mount. Read only permissions for creating local backups and/or archives
+Python script to collect all the drive_ids in sharepoint associated to a team and create a combine mount. Read only permissions for creating local backups and/or archives
 
 ## Why?
 
@@ -13,9 +13,8 @@ Becuase microsoft dont like to make things easy and create sharepoint subsites f
 3. Once the app has been made make note of the Application (client) ID and Directory (tennant) ID
 4. Go to Certificates & Secrets > Client secrets and create a new client secret, set the expire and click add
 5. Once it has been generated make a note of the Value, once you change page that will be stared out and you wont be able to see it again.
-6. In your config.json set the `tennantId`, `clientID` and `client secret (Value)`
-7. in the remote-template.conf set the `client ID`, `client secret (value)`, `tennantID` and replace `<tenant>` with your tenantID
-8. Go to `API permissions > Add a permission > Microsoft Graph > Application permissions` and add the below permissions
+6. In your config.json set the `tenantId`, `clientID` and `clientSecret` (Value)
+7. Go to `API permissions > Add a permission > Microsoft Graph > Application permissions` and add the below permissions
 
 ```plain
 Channel.ReadBasic.All
@@ -26,16 +25,29 @@ Sites.Read.All
 Team.ReadBasic.All
 ```
 
-9. Once those have been added go to `Add a permission > Sharepoint  > Application permissions` and add the below permission
+8. Once those have been added go to `Add a permission > Sharepoint  > Application permissions` and add the below permission
 
 ```plain
 Sites.Read.All
 ```
 
+9. Grant admin consent.
+
+## Flags
+
+```txt
+--team-name   [TEAM_NAME ...]
+              Team display name prefix(es). Required unless --team-id
+--team-id     Team id (skips name lookup). Required unless --team-name
+--exact       Exact displayName match
+--config      Path to config.json
+-o, --outdir  Output directory for .conf files
+-h, --help    Show help
+```
+
 ## Usage
 
-- Run the script
-- Enter your team name when it asks
+- Make sure you have python 3.7+ and rclone in your PATH
+- Run `python3 generate_config.py --team-name "Team One" "Team Two"`
 - Wait
-- Run `rclone lsd --config ./rclone TEAM_NAME_union:` when the script is finished to confirm you can see everything
-- If the team only contains standard channels it will create a config with 1 remote
+- Run `rclone lsd --config ./Team_One.conf Team_One_combine:` when the script is finished to confirm you can see everything
diff --git a/T2R.ps1 b/T2R.ps1
deleted file mode 100644
index 514551f..0000000
--- a/T2R.ps1
+++ /dev/null
@@ -1,137 +0,0 @@
-param(
-    [Parameter(Mandatory = $true)]
-    [string[]]$Team,
-    [switch]$Exact
-)
-
-$config = Get-Content -Path ./config.json -Raw | ConvertFrom-Json
-$tenantId = $config.tenantId
-$clientID = $config.clientID
-$clientSecret = (ConvertTo-SecureString $config.clientSecret -AsPlainText -Force )
-$Scope = "https://graph.microsoft.com/.default"
-$authToken = Get-MsalToken -ClientId $clientID -ClientSecret $clientSecret -TenantId $tenantId -Scopes $Scope
-
-Write-Host "Processing $($Team.Count) teams: $($Team -join ', ')"
-
-$Headers = @{
-    "Authorization" = "Bearer $($authToken.AccessToken)"
-    "Content-type"  = "application/json"
-}
-
-$allTeams = @()
-$teamsResponse = Invoke-WebRequest -Uri "https://graph.microsoft.com/v1.0/teams/" -Headers $Headers -Method GET
-$aadTeams = $teamsResponse.Content | ConvertFrom-Json
-$allTeams += $aadTeams.value
-if ($aadTeams.'@odata.nextLink') {
-    do {
-        $teamsResponse = Invoke-WebRequest -Uri $aadTeams.'@odata.nextLink' -Headers $Headers -Method GET
-        $aadTeams = $teamsResponse.Content | ConvertFrom-Json
-        $allTeams += $aadTeams.value
-    } until (!$aadTeams.'@odata.nextLink')
-}
-$aadTeams = $allTeams
-
-foreach ($requestedTeam in $Team) {
-    Write-Host "`nProcessing team: $requestedTeam"
-    
-    $channels = [System.Collections.ArrayList]::New()
-    $foundTeam = $false
-    
-    foreach ($teams in $aadTeams) {
-        if ($foundTeam) { break }
-        
-        $teamMatches = if ($Exact) {
-            $teams.displayName -eq $requestedTeam
-        } else {
-            $teams.displayName -like "$($requestedTeam)*"
-        }
-        
-        if ($teamMatches) {
-            Write-Host "Found matching team: $($teams.displayName)"
-            $apiUri = "https://graph.microsoft.com/v1.0/teams/$($teams.id)/channels"
-            
-            $response = Invoke-WebRequest -Headers $Headers -Uri $apiUri -Method GET
-            
-            try {
-                $teamResponse = $response.Content | ConvertFrom-Json -ErrorAction Stop
-            } catch {
-                $teamResponse = [System.Web.HttpUtility]::JavaScriptStringDeserializer.Deserialize($response.Content, [System.Object])
-            }
-            
-            $channelList = $null
-            
-            if ($teamResponse.value) {
-                $channelList = $teamResponse.value
-            } elseif ($teamResponse.PSObject.Properties['value']) {
-                $channelList = $teamResponse.PSObject.Properties['value'].Value
-            } else {
-                $rawContent = $response.Content
-                if ($rawContent -match '"value":\s*\[(.*?)\]') {
-                    $jsonArray = "[$($matches[1])]"
-                    $channelList = $jsonArray | ConvertFrom-Json
-                }
-            }
-            
-            if ($channelList) {
-                Write-Host "Processing $($channelList.Count) channels for team: $($teams.displayName)"
-                foreach ($channel in $channelList) {
-                    Start-Sleep -Milliseconds 250
-                    $folderResponse = Invoke-WebRequest -Headers $Headers -Uri "https://graph.microsoft.com/v1.0/teams/$($teams.id)/channels/$($channel.id)/filesFolder" -Method GET
-                    $channelFolder = $folderResponse.Content | ConvertFrom-Json
-                    $DriveID = $channelFolder.parentReference.driveId
-                    $displayName = $channelFolder.name -replace " ", "_" -replace ",", ""
-                    $Result = [PSCustomObject]@{
-                        displayName    = $displayName
-                        driveID        = $DriveID
-                        membershipType = $channel.membershipType
-                        teamName       = $teams.displayName
-                    }
-                    [void]$channels.Add($Result)
-                    Write-Host "Added channel: $displayName with DriveID: $DriveID"
-                }
-            } else {
-                Write-Host "WARNING: No channels found for team: $($teams.displayName)"
-            }
-            $foundTeam = $true
-        }    
-    }
-    
-    if ($channels.Count -eq 0) {
-        Write-Host "ERROR: No channels were processed for team: $requestedTeam"
-        continue
-    }
-    
-    Write-Host "Total channels processed for $requestedTeam`: $($channels.Count)"
-    
-    $uniqueDrives = $channels | Sort-Object -Property driveID -Unique | Select-Object displayName, driveID, membershipType, teamName
-    Write-Host "Unique drives found for $requestedTeam`: $($uniqueDrives.Count)"
-    
-    $uniqueDrives | ForEach-Object { 
-        if ($_.membershipType -eq 'standard') { 
-            $_.displayName = $_.teamName -replace " ", "_" 
-        }
-    }
-    
-    $UnionName = "$($channels[0].teamName)_union" -replace " ", "_"
-    $RcloneFileName = $UnionName -replace "_union", ""
-    
-    foreach ($drive in $uniqueDrives) {
-        $config = Get-Content -Path ./remote-template.conf -Raw
-        $remote = $config -f $drive.displayName, $drive.driveID
-        Add-Content -Value $remote -Path "./$RcloneFileName.conf"
-        Write-Host "Added remote config for: $($drive.displayName)"
-    }
-    
-    if ($uniqueDrives.Count -ge 2) {
-        $teamNames = $($uniqueDrives.displayName -join [Environment]::NewLine) -replace "\n", ": "
-        $teamNames = "$teamNames" + ":"
-        $union = Get-Content -Path ./union-template.conf -Raw
-        $unionConf = $union -f $UnionName, $teamNames
-        Add-Content -Value $unionConf -Path "./$RcloneFileName.conf"
-        Write-Host "Generated union config: $UnionName"
-    } else {
-        Write-Host "Skipping union creation - only $($uniqueDrives.Count) remote(s) found (union requires 2+)"
-    }
-    
-    Write-Host "Generated config file: $RcloneFileName.conf"
-}
\ No newline at end of file
diff --git a/combine-template.conf b/combine-template.conf
new file mode 100644
index 0000000..dce9c84
--- /dev/null
+++ b/combine-template.conf
@@ -0,0 +1,3 @@
+[{0}]
+type = combine
+upstreams = {1}
diff --git a/generate_config.py b/generate_config.py
new file mode 100644
index 0000000..bba424e
--- /dev/null
+++ b/generate_config.py
@@ -0,0 +1,206 @@
+#!/usr/bin/env python3
+"""Generate rclone combine configs for Microsoft Teams channel drives."""
+
+from __future__ import annotations
+
+import argparse
+import json
+import re
+import sys
+import time
+import urllib.error
+import urllib.parse
+import urllib.request
+from pathlib import Path
+
+GRAPH = "https://graph.microsoft.com/v1.0"
+HERE = Path(__file__).resolve().parent
+DEFAULT_CONFIG = HERE / "config.json"
+REMOTE_TMPL = (HERE / "remote-template.conf").read_text()
+COMBINE_TMPL = (HERE / "combine-template.conf").read_text()
+
+
+def sanitize_remote(name: str) -> str:
+    return re.sub(r"[^\w.-]+", "_", name).strip("_")
+
+
+def get_token(tenant: str, client_id: str, client_secret: str) -> str:
+    data = urllib.parse.urlencode(
+        {
+            "client_id": client_id,
+            "client_secret": client_secret,
+            "scope": "https://graph.microsoft.com/.default",
+            "grant_type": "client_credentials",
+        }
+    ).encode()
+    req = urllib.request.Request(
+        f"https://login.microsoftonline.com/{tenant}/oauth2/v2.0/token",
+        data=data,
+        method="POST",
+    )
+    with urllib.request.urlopen(req) as resp:
+        return json.load(resp)["access_token"]
+
+
+def graph_get(url: str, token: str) -> dict:
+    req = urllib.request.Request(url, headers={"Authorization": f"Bearer {token}"})
+    with urllib.request.urlopen(req) as resp:
+        return json.load(resp)
+
+
+def graph_paged(url: str, token: str) -> list:
+    items: list = []
+    while url:
+        payload = graph_get(url, token)
+        items.extend(payload.get("value", []))
+        url = payload.get("@odata.nextLink")
+    return items
+
+
+def find_team(teams: list, name: str, exact: bool) -> dict | None:
+    for team in teams:
+        dn = team.get("displayName") or ""
+        if exact and dn == name:
+            return team
+        if not exact and dn.lower().startswith(name.lower()):
+            return team
+    return None
+
+
+def collect_channels(team_id: str, token: str) -> list[dict]:
+    channels = graph_paged(f"{GRAPH}/teams/{team_id}/channels", token)
+    results = []
+    for ch in channels:
+        time.sleep(0.25)
+        folder = graph_get(
+            f"{GRAPH}/teams/{team_id}/channels/{ch['id']}/filesFolder", token
+        )
+        results.append(
+            {
+                "displayName": ch["displayName"],
+                "driveId": folder["parentReference"]["driveId"],
+                "membershipType": ch.get("membershipType", "standard"),
+            }
+        )
+        print(f"  {ch['membershipType']}: {ch['displayName']} -> {folder['parentReference']['driveId']}")
+    return results
+
+
+def build_config(team_name: str, channels: list[dict], creds: dict) -> str:
+    team_remote = sanitize_remote(team_name)
+    standard = [c for c in channels if c["membershipType"] == "standard"]
+    other = [c for c in channels if c["membershipType"] != "standard"]
+
+    blocks: list[str] = []
+    upstreams: list[str] = []
+    written_remotes: set[str] = set()
+
+    def remote_block(name: str, drive_id: str) -> str:
+        return REMOTE_TMPL.format(
+            name=name,
+            drive_id=drive_id,
+            client_id=creds["client_id"],
+            client_secret=creds["client_secret"],
+            tenant=creds["tenant"],
+        )
+
+    if standard:
+        blocks.append(remote_block(team_remote, standard[0]["driveId"]))
+        written_remotes.add(team_remote)
+        for ch in standard:
+            name = ch["displayName"]
+            upstreams.append(f'"{name}={team_remote}:{name}"')
+
+    for ch in other:
+        remote = sanitize_remote(ch["displayName"])
+        if remote in written_remotes:
+            remote = f"{remote}_{ch['membershipType']}"
+        blocks.append(remote_block(remote, ch["driveId"]))
+        written_remotes.add(remote)
+        name = ch["displayName"]
+        # private drives nest files under a folder named after the channel
+        if ch["membershipType"] == "private":
+            upstreams.append(f'"{name}={remote}:{name}"')
+        else:
+            # shared / unknownFutureValue: files sit at drive root
+            upstreams.append(f'"{name}={remote}:"')
+
+    if upstreams:
+        blocks.append(COMBINE_TMPL.format(f"{team_remote}_combine", " ".join(upstreams)))
+
+    return "\n".join(blocks)
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--team-name", nargs="+", help="Team display name prefix(es)")
+    parser.add_argument("--team-id", help="Team id (skips name lookup)")
+    parser.add_argument("--exact", action="store_true", help="Exact displayName match")
+    parser.add_argument(
+        "--config",
+        type=Path,
+        default=DEFAULT_CONFIG,
+        help=f"Path to config.json (default: {DEFAULT_CONFIG})",
+    )
+    parser.add_argument(
+        "-o",
+        "--outdir",
+        type=Path,
+        default=HERE,
+        help="Output directory for .conf files",
+    )
+    args = parser.parse_args()
+
+    if not args.team_id and not args.team_name:
+        parser.error("provide --team-name and/or --team-id")
+
+    cfg = json.loads(args.config.read_text())
+    creds = {
+        "tenant": cfg["tenantId"],
+        "client_id": cfg["clientID"],
+        "client_secret": cfg["clientSecret"],
+    }
+
+    print("Authenticating...")
+    token = get_token(creds["tenant"], creds["client_id"], creds["client_secret"])
+    args.outdir.mkdir(parents=True, exist_ok=True)
+
+    targets = []
+    if args.team_id:
+        team = graph_get(f"{GRAPH}/teams/{args.team_id}", token)
+        print(f"Team: {team['displayName']} ({args.team_id})")
+        targets.append(team)
+
+    if args.team_name:
+        print("Listing teams...")
+        all_teams = graph_paged(f"{GRAPH}/teams/", token)
+        for requested in args.team_name:
+            team = find_team(all_teams, requested, args.exact)
+            if not team:
+                print(f"ERROR: no team matched '{requested}'", file=sys.stderr)
+                continue
+            print(f"Found: {team['displayName']} ({team['id']})")
+            targets.append(team)
+
+    for team in targets:
+        print(f"\nChannels for {team['displayName']}:")
+        try:
+            channels = collect_channels(team["id"], token)
+        except urllib.error.HTTPError as e:
+            print(f"ERROR fetching channels: {e}", file=sys.stderr)
+            continue
+
+        if not channels:
+            print(f"ERROR: no channels for {team['displayName']}", file=sys.stderr)
+            continue
+
+        conf = build_config(team["displayName"], channels, creds)
+        out = args.outdir / f"{sanitize_remote(team['displayName'])}.conf"
+        out.write_text(conf)
+        print(f"Wrote {out}")
+
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
diff --git a/remote-template.conf b/remote-template.conf
index 38e77fc..f80e76a 100644
--- a/remote-template.conf
+++ b/remote-template.conf
@@ -1,12 +1,12 @@
-[{0}]
+[{name}]
 type = onedrive
-drive_id = {1}
+drive_id = {drive_id}
 drive_type = documentLibrary
-client_id = 
-client_secret = 
-tenant = 
+client_id = {client_id}
+client_secret = {client_secret}
+tenant = {tenant}
 client_credentials = true
 access_scopes = Files.Read Files.Read.All Sites.Read.All offline_access
 link_scope = organization
-auth_url = https://login.microsoftonline.com/<tenant>/oauth2/v2.0/authorize
-token_url = https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token
+auth_url = https://login.microsoftonline.com/{tenant}/oauth2/v2.0/authorize
+token_url = https://login.microsoftonline.com/{tenant}/oauth2/v2.0/token
diff --git a/union-template.conf b/union-template.conf
deleted file mode 100644
index 5920179..0000000
--- a/union-template.conf
+++ /dev/null
@@ -1,3 +0,0 @@
-[{0}]
-type = union
-upstreams = {1}
\ No newline at end of file