fix: load paths, imports, transpile, and CI permissions

.github/workflows/lint_sqlfluff.yml

@@ -35,8 +35,12 @@ jobs:
         shell: bash
         run: sqlfluff lint --format github-annotation --annotation-level failure --nofail ${{ steps.get_files_to_lint.outputs.lintees }} > annotations.json
       - name: Annotate
+        # Fork PRs cannot create check runs with GITHUB_TOKEN; lint already ran.
+        continue-on-error: true
+        if: steps.get_files_to_lint.outputs.lintees != ''
         uses: yuzutech/annotations-action@v0.6.0
         with:
           repo-token: "${{ secrets.GITHUB_TOKEN }}"
           title: "SQLFluff Lint"
-          input: "./annotations.json"
+          input: "./annotations.json"
+          ignore-missing-file: true

mimic-iii/buildmimic/duckdb/import_duckdb.sh

@@ -29,7 +29,7 @@ usage () {
     die "
 USAGE: ./import_duckdb.sh mimic_data_dir [output_db]
 WHERE:
-    mimic_data_dir        directory that contains csv.gz or csv files
+    mimic_data_dir        directory that contains csv.tar.gz or csv files
     output_db: optional   filename for duckdb file (default: mimic3.db)\
 "
 }
@@ -84,9 +84,11 @@ make_table_name () {
 # load data into database
 find "$MIMIC_DIR" -type f -regex '.*\.csv\(.gz\)*' | while IFS= read -r FILE; do
     make_table_name "$FILE"
-    echo "Loading $FILE .. \c"
+    # Escape single quotes for SQL string literal
+    FILE_SQL=$(printf '%s' "$FILE" | sed "s/'/''/g")
+    printf "Loading %s .. " "$FILE"
     try duckdb "$OUTFILE" <<-EOSQL
-		COPY $TABLE_NAME FROM '$FILE' (HEADER, DELIM ',', QUOTE '"', ESCAPE '"');
+		COPY $TABLE_NAME FROM '$FILE_SQL' (HEADER, DELIM ',', QUOTE '"', ESCAPE '"');
 EOSQL
     echo "done!"
 done && echo "Successfully finished loading data into $OUTFILE."

mimic-iii/buildmimic/postgres/Makefile

@@ -104,7 +104,7 @@ mimic-build-gz:
 	@echo '------------------'
 	@echo ''
 	@sleep 2
-	psql "$(DBSTRING)" -v ON_ERROR_STOP=1 -f postgres_load_data_gz.sql -v mimic_data_dir=${datadir}
+	psql "$(DBSTRING)" -v ON_ERROR_STOP=1 -f postgres_load_data_gz.sql -v "mimic_data_dir=$(DATADIR)"
 	@echo ''
 	@echo '--------------------'
 	@echo '-- Adding indexes --'
@@ -151,7 +151,7 @@ mimic-build:
 	@echo '------------------'
 	@echo ''
 	@sleep 2
-	psql "$(DBSTRING)" -v ON_ERROR_STOP=1 -f postgres_load_data.sql -v mimic_data_dir=${DATADIR}
+	psql "$(DBSTRING)" -v ON_ERROR_STOP=1 -f postgres_load_data.sql -v "mimic_data_dir=$(DATADIR)"
 	@echo ''
 	@echo '--------------------'
 	@echo '-- Adding indexes --'
@@ -184,15 +184,15 @@ ifeq ("$(physionetuser)","")
 	@echo 'Call the makefile again with physionetuser=<USERNAME>'
 	@echo ' e.g. make eicu-download datadir=/path/to/data physionetuser=hello@physionet.org'
 else
-	wget --user $(physionetuser) --ask-password -P $(DATADIR) -A csv.gz -m -p -E -k -K -np -nd "$(PHYSIONETURL)"
+	wget --user $(physionetuser) --ask-password -P "$(DATADIR)" -A csv.gz -m -p -E -k -K -np -nd "$(PHYSIONETURL)"
 endif
 
 mimic-demo-download:
 	@echo '------------------------------------------'
 	@echo '-- Downloading MIMIC-III from PhysioNet --'
 	@echo '------------------------------------------'
 	@echo ''
-	wget --user $(physionetuser) --ask-password -P $(DATADIR) -A csv.gz -m -p -E -k -K -np -nd "$(PHYSIONETDEMOURL)"
+	wget --user $(physionetuser) --ask-password -P "$(DATADIR)" -A csv.gz -m -p -E -k -K -np -nd "$(PHYSIONETDEMOURL)"
 
 #This is fairly inelegant and could be tidied with a for loop and an if to check for gzip,
 #but need to maintain compatibility with Windows, which baffling lacks these things
@@ -281,7 +281,7 @@ concepts:
 	@echo '---------------------'
 	@echo ''
 	@sleep 2
-	cd ../../concepts_postgres/ && psql "$(DBSTRING)" -v ON_ERROR_STOP=1 -c "CREATE SCHEMA IF NOT EXISTS mimiciii_derived;" -f postgres-make-concepts.sql
+	cd ../../concepts/ && psql "$(DBSTRING)" -v ON_ERROR_STOP=1 -f make-concepts.sql
 
 
 .PHONY: help mimic clean

mimic-iii/buildmimic/postgres/create_mimic_user.sh

@@ -20,6 +20,13 @@ else
   echo "User is set to '$MIMIC_USER'";
 fi
 
+# escape ' in password for SQL string literal
+MIMIC_PASSWORD_SQL=$(printf '%s' "$MIMIC_PASSWORD" | sed "s/'/''/g")
+# quote SQL identifiers (double any embedded ")
+sql_ident () { printf '%s' "$1" | sed 's/"/""/g; s/^/"/; s/$/"/'; }
+MIMIC_USER_SQL=$(sql_ident "$MIMIC_USER")
+MIMIC_DB_SQL=$(sql_ident "$MIMIC_DB")
+
 PSQL='psql'
 
 # add in the host/port, if they were specified (not null, -n)
@@ -58,11 +65,11 @@ fi
 if [ "$MIMIC_USER" != "postgres" ]; then
     # we need to create this user via postgres
     # use SUDO to login as postgres
-    $PSQL -U postgres -d postgres -c "DROP USER IF EXISTS $MIMIC_USER; CREATE USER $MIMIC_USER WITH PASSWORD '$MIMIC_PASSWORD';"
+    $PSQL -U postgres -d postgres -c "DROP USER IF EXISTS $MIMIC_USER_SQL; CREATE USER $MIMIC_USER_SQL WITH PASSWORD '$MIMIC_PASSWORD_SQL';"
 fi
 
 if [ "$MIMIC_DB" != "postgres" ]; then
   # drop and recreate the database
-  $PSQL -U postgres -d postgres -c "DROP DATABASE IF EXISTS $MIMIC_DB;"
-  $PSQL -U postgres -d postgres -c "CREATE DATABASE $MIMIC_DB OWNER $MIMIC_USER;"
-fi
+  $PSQL -U postgres -d postgres -c "DROP DATABASE IF EXISTS $MIMIC_DB_SQL;"
+  $PSQL -U postgres -d postgres -c "CREATE DATABASE $MIMIC_DB_SQL OWNER $MIMIC_USER_SQL;"
+fi

mimic-iii/buildmimic/sqlite/import.py

@@ -10,6 +10,34 @@
 CHUNKSIZE = 10 ** 6
 CONNECTION_STRING = "sqlite:///{}".format(DATABASE_NAME)
 
+# Column dtypes for tables that trigger pandas mixed-type warnings when
+# loaded with the default low_memory chunked inference (see #1237).
+# Types mirror mimic-iii/buildmimic/postgres/postgres_create_tables.sql.
+TABLE_DTYPES = {
+    "chartevents": {
+        "WARNING": "Int64",
+        "ERROR": "Int64",
+        "RESULTSTATUS": "string",
+        "STOPPED": "string",
+    },
+    "datetimeevents": {
+        "WARNING": "Int64",
+        "ERROR": "Int64",
+        "RESULTSTATUS": "string",
+        "STOPPED": "string",
+    },
+    "inputevents_cv": {
+        "ORIGINALRATE": "float64",
+        "ORIGINALRATEUOM": "string",
+        "ORIGINALSITE": "string",
+    },
+    "noteevents": {
+        "CHARTTIME": "string",
+        "STORETIME": "string",
+        "ISERROR": "string",
+    },
+}
+
 
 def _table_name_from_csv(filename: str) -> str:
     """Derive SQL table name from a CSV path (literal suffix, not str.strip)."""
@@ -21,20 +49,39 @@ def _table_name_from_csv(filename: str) -> str:
     return name.lower()
 
 
+def _read_csv(path, table, **kwargs):
+    # low_memory=False avoids dtype re-inference across chunks; table-specific
+    # dtypes pin the columns that otherwise flip between numeric/object.
+    return pd.read_csv(
+        path,
+        index_col="ROW_ID",
+        low_memory=False,
+        dtype=TABLE_DTYPES.get(table),
+        **kwargs,
+    )
+
+
 if os.path.exists(DATABASE_NAME):
     msg = "File {} already exists.".format(DATABASE_NAME)
     print(msg)
     sys.exit()
 
+# Prefer .csv.gz when both exist for the same table; also load plain .csv
+# (import.sh already accepts both).
+files_by_table = {}
+for f in glob("*.csv"):
+    files_by_table[_table_name_from_csv(f)] = f
 for f in glob("*.csv.gz"):
+    files_by_table[_table_name_from_csv(f)] = f
+
+for table, f in sorted(files_by_table.items()):
     print("Starting processing {}".format(f))
-    table = _table_name_from_csv(f)
     if os.path.getsize(f) < THRESHOLD_SIZE:
-        df = pd.read_csv(f, index_col="ROW_ID")
+        df = _read_csv(f, table)
         df.to_sql(table, CONNECTION_STRING)
     else:
         # If the file is too large, let's do the work in chunks
-        for chunk in pd.read_csv(f, index_col="ROW_ID", chunksize=CHUNKSIZE):
+        for chunk in _read_csv(f, table, chunksize=CHUNKSIZE):
             chunk.to_sql(table, CONNECTION_STRING, if_exists="append")
     print("Finished processing {}".format(f))
 

mimic-iii/buildmimic/sqlite/import.sh

@@ -28,19 +28,19 @@ for FILE in *; do
     TABLE_NAME=$(echo "${FILE%%.*}" | tr "[:upper:]" "[:lower:]")
     case "$FILE" in
         *csv)
-            IMPORT_CMD=".import $FILE $TABLE_NAME"
+            IMPORT_CMD=".import \"$FILE\" $TABLE_NAME"
         ;;
         # need to decompress csv before load
         *csv.gz)
-            IMPORT_CMD=".import \"|gzip -dc $FILE\" $TABLE_NAME"
+            IMPORT_CMD=".import \"|gzip -dc \\\"$FILE\\\"\" $TABLE_NAME"
         ;;
         # not a data file so skip
         *)
             continue
         ;;
     esac
     echo "Loading $FILE."
-    sqlite3 $OUTFILE <<EOF
+    sqlite3 "$OUTFILE" <<EOF
 .headers on
 .mode csv
 $IMPORT_CMD