DHI.Services.MIKECloud for Documents — Internal Developer Guide

Use MIKECloud as the backend for the Documents domain (flat and grouped). For platform/auth/path rules, see MIKECloud Core.

---

What you get

Two flavors sharing the same concepts:

• GroupedDocumentRepository (BaseGroupedDocumentRepository<string>) Best for most usage: groups/folders, group filters, metadata filtering.
• DocumentRepository (IDocumentRepository<string>) Flat repo. No group APIs, but accepts full-path IDs.

Mapping to MIKECloud

Documents concept	MIKECloud object
Group/folder path	Project + nested Subprojects
Document (payload)	File dataset (DatasetType.File)
Document metadata (Parameters)	Dataset Metadata (string values)
Document ID	Full path /…/…/name.ext

---

Capabilities & differences

Operation / Area	Grouped repo	Flat repo	Notes
Add (upload)	✓	✓	Creates missing subprojects. Upload via Transfer; then copies Parameters → Metadata
Get (download)	✓	✓	Returns (Stream, fileType, fileName); Grouped rewinds stream, Flat doesn’t
Remove	✓	✓	Hard-deletes dataset
Contains	✓	✓	Resolves dataset by full path
Count	✓	✓	Counts File datasets recursively
GetIds	✓	✓	Returns full path IDs
GetAll	✓	✗	Emits Document<string> (Id/Name/Group + Metadata)
GetByGroup	✓	✗	Exact group match (normalized, case-insensitive)
GetMetadata	✓	✓	Dictionary from dataset Metadata
GetAllMetadata / Filter	✓	✗	Grouped supports value-token AND filtering

---

Paths & IDs

• Full path ID: "group/full/path/FileName.ext".
• The path (before the last /) → subprojects (created on demand by Add).
• The last segment → dataset/file name in MIKECloud.

---

Connecting

Connection string

apiKey=<APIKEY>;projectId=<PROJECTID>;environment=Prod

Direct (code)

var grouped = new DHI.Services.Provider.MIKECloud.GroupedDocumentRepository(connString);
var flat    = new DHI.Services.Provider.MIKECloud.DocumentRepository(connString);

groupedSvc = new new GroupedDocumentService<string>(grouped);
flatSvc = new new DocumentService<string>(flat);

ServiceLocator.Register(groupedSvc, "grouped-mc-docs");
ServiceLocator.Register(flatSvc, "flat-mc-docs");

Web API (Connections module)

{
  "$type": "DHI.Services.Documents.WebApi.GroupedDocumentServiceConnection, DHI.Services.Documents.WebApi",
  "RepositoryType": "DHI.Services.Provider.MIKECloud.GroupedDocumentRepository, DHI.Services.Provider.MIKECloud",
  "ConnectionString": "apiKey=[env:MIKECLOUD_APIKEY];projectId=[env:MIKECLOUD_PROJECT];environment=Prod",
  "Name": "Documents via MIKECloud (grouped)",
  "Id": "mc-docs"
}

---

Code recipes

Upload + metadata → download

var docs = new GroupedDocumentRepository(conn);
using var file = File.OpenRead(@"C:\Temp\Summary.pdf");
var meta = new Parameters { { "Department","Finance" }, { "Confidential","Yes" } };

docs.Add(file, "reports/2025/Q3/Summary.pdf", meta);

var (stream, ext, name) = docs.Get("reports/2025/Q3/Summary.pdf");
// Grouped returns a rewound stream
using var outFs = File.Create($@"C:\Temp\{name}");
stream.CopyTo(outFs);

Flat repo download (remember to rewind)

var flat = new DocumentRepository(conn);
var (s, ext, name) = flat.Get("/Models/HEC/v2/notes.txt");
s.Seek(0, SeekOrigin.Begin);

---

Troubleshooting & notes

• Project not found: constructor validates projectId → ArgumentException.
• Remove “Invalid path …” → Provided path is not a dataset; ensure the last segment is a file name.
• Get: fileType includes the leading dot (e.g., .csv).
• Large files: Upload/download use the platform TransferClient (streamed); avoid buffering.
• Dataset overwrite: Re-uploading the same name in the same folder updates the existing dataset per platform rules.