Merge pull request #38 from PhilanthropyDataCommons/1-add-architecture

Add architecture documentation

Author: slifty

docs/ARCHITECTURE.md

@@ -0,0 +1,52 @@
+
+# Architecture
+
+## Diagram
+
+(This diagram is generated by [mermaid](https://mermaid-js.github.io/mermaid/#/./flowchart?id=flowcharts-basic-syntax))
+
+```mermaid
+graph LR
+
+    subgraph GMS Access
+        GMS_APIs((GMS APIs))
+        GMS_Dumps[(GMS Data Dumps)]
+    end
+    subgraph Client Applications
+        Scraper[[GMS Scrapers]]
+        Ingester[[GMS Ingesters]]
+        UX{{User Interfaces}}
+        GMS[Integrated GMSs]
+        Infomediaries[[Infomediaries]]
+    end
+    subgraph PDC Backend
+        API((API))
+        Database[(Data Store)]
+    end
+    subgraph External Services
+        GMS_APIs((GMS APIs))
+        GMS_Dumps[(GMS Data Dumps)]
+        Auth((KeyCloak))
+        Email((MailGun))
+    end
+
+    GMS---API
+    Infomediaries---API
+    Scraper---GMS_APIs
+    Scraper---API
+    Ingester---API
+    Ingester---GMS_Dumps
+    API---Database
+    API---UX
+    API---Auth
+    API---Email
+```
+
+## Concepts
+
+The core of the PDC is its API.  All third party integrations and user interfaces are powered by that API.
+
+* `GMS Scrapers` pull data from a GMS api on a regular basis and inject that data into the PDC.
+* `GMS Ingesters` are provided data by a GMS (e.g. in the form of batch exports) and inject that data into the PDC.
+* `Integrated GMSs` are GMSs that know how to inject data into the PDC directly.
+* `Infomediaries` are third party information sources, and might be scraped / ingested (TBD based on what we learn).