Add schema IDs to glean merged files

Summary:
Ensure that all fact files merged using `glean merge` have a schema id; and that ID is the same across all input and output files.

## Part 1)
Support consistent Schema IDs when importing a facts JSON into an existing DB, which has the autoritativet schema to align with.
NOTE: This approach relies on JSON fact files implementing the `schema_id` metadata attribute, as implemented in T214992259. This limits which indexers' JSON fact files can be merged to those that implement `schema_id`. At the moment, C++ is OK, for example, but Python is not.

## Part 2)
Merge multiple JSON files into a single JSON file type merge using the inventory mechanism

# This diff implements...
This diff adds schema ID assertions to implement  Part 1)

Reviewed By: pepeiborra

Differential Revision: D71329107

fbshipit-source-id: e7276e2546edf11a41ace05eb3f69738d95d9cc6

Author: simonhollis

glean/tools/gleancli/GleanCLI/Merge.hs

@@ -22,6 +22,7 @@ import System.Process
 
 import Control.Concurrent.Stream
 import Util.OptParse
+import Util.Log
 import Thrift.Protocol.Compact
 
 import Glean.LocalOrRemote (loadDbSchema)
@@ -41,6 +42,7 @@ import Glean.Write (fileToBatches, schemaIdToOpts)
 import Glean.Write.JSON (buildJsonBatch)
 import System.Directory.Extra (listFiles)
 
+
 data MergeCommand = MergeCommand
   { mergeFiles :: [FilePath]
   , mergeFileSize :: Int
@@ -82,6 +84,7 @@ instance Plugin MergeCommand where
         dbSchema <- Glean.withBackendWithDefaultOptions
           _evb _cfgAPI _svc Nothing $ \backend -> do
             loadDbSchema backend repo
+        logInfo("db's schema ID is: "  <> show(schemaId dbSchema))
         return (schemaInventory dbSchema, Just dbSchema)
       Right mergeInventory -> do
         inventory <- Inventory.deserialize <$> B.readFile mergeInventory
@@ -141,16 +144,30 @@ instance Plugin MergeCommand where
         where
         read :: FilePath -> Int -> FactSet -> IO FactOwnership
         read file size factSet = do
-          hPutStrLn stderr $ "Reading " <> file <> " (" <> show size <> ")"
+          logInfo $ "Reading " <> file <> " (" <> show size <> " bytes)"
           batch <- case fileFormat of
             JsonFormat -> do
               case dbSchema of
                 Nothing -> throwIO $ ErrorCall $
                   "No db schema to serialize json format file. "
                   <> "Please specify the database"
                 Just schema -> do
-                  (batches, schema_id) <- fileToBatches file
-                  buildJsonBatch schema (schemaIdToOpts schema_id) batches
+                  (batches, schema_id_file) <- fileToBatches file
+                  if Just(schemaId schema) == schema_id_file then
+                    logInfo(
+                      "Schema matches with db schema. Merging data from "
+                      <>  file
+                      )
+                  else
+                    throwIO $ ErrorCall $
+                        "ERROR - ABORTING MERGE\nSchema ID mismatch:\ndb: "
+                        <> show(schemaId schema) <> "\nvs\nFile: "
+                        <> file <> " has " <> show schema_id_file
+                  let getSchemaId theschema = Just(schemaId theschema) in
+                    buildJsonBatch schema
+                        (schemaIdToOpts $ getSchemaId schema) batches
+
+
             BinaryFormat -> do
               bytes <- B.readFile file
               case deserializeCompact bytes of