add ability to explicitly specify additional mavedb scoresets matched with uniprot ids

1 files changed, 25 insertions, 12 deletions

diff --git a/src/scrape.hs b/src/scrape.hs
index 56f4631..e877a9c 100644
--- a/src/scrape.hs
+++ b/src/scrape.hs
@@ -6,6 +6,8 @@
 module Main where
 
 import Control.Lens
+import Control.Monad
+import Data.Aeson.Lens
 import Data.List
 import qualified Data.Map as M
 import Data.Maybe
@@ -32,30 +34,41 @@ schema =
       PRIMARY KEY (gene, hgvs_p));
   |]
 
-data Config = Config FilePath Double
+data Config = Config FilePath Double (Maybe FilePath)
 
 main = configured =<< execParser opts
   where
     config =
       Config
         <$> strArgument (metavar "DB" <> help "path to SQLite DB to insert into")
-        <*> option
-          auto
-          ( long "min-pi"
-              <> short 'p'
-              <> metavar "π"
-              <> help "minimum estimated labelling proportion"
-              <> showDefault
-              <> value 0.1
-          )
+          <*> option
+            auto
+            ( long "min-pi"
+                <> short 'p'
+                <> metavar "π"
+                <> help "minimum estimated labelling proportion"
+                <> showDefault
+                <> value 0.1
+            )
+          <*> option
+            auto
+            ( long "extra"
+                <> short 'e'
+                <> metavar "PATH"
+                <> help "path to JSON map between URNs and uniprot ids to supplement MaveDB URN query"
+                <> value Nothing
+            )
 
     opts =
       info
         (config <**> helper)
         (fullDesc <> progDesc "scrapes MaveDB for human proteins, normalises, and inserts HGVS notation into a SQLite DB" <> header "scrape -- MaveDB scraper")
 
-configured (Config out minpi) = do
+configured (Config out minpi uniprotMap) = do
   urns <- queryURNs
+  extraUrns <- forM uniprotMap $ \path -> do
+    json <- T.readFile path
+    return $ zip (json ^.. members . asIndex) (json ^.. members . _String)
   scores <-
     M.fromListWith (M.unionWith (++)) . catMaybes
       <$> mapM
@@ -65,7 +78,7 @@ configured (Config out minpi) = do
             s <- normalise minpi <$> getScores x
             pure $ (g,) <$> s
         )
-        urns
+        (urns ++ fromMaybe mempty extraUrns)
 
   let cScores = M.map (M.toList . M.map geomean) scores