Validate signal names before performing API calls (#47)

* Validate signal names before performing API calls

VERSION

@@ -1,2 +1 @@
-0.0.25
-
+0.0.26

exabel_data_sdk/scripts/load_time_series_from_csv.py

@@ -1,14 +1,18 @@
 import argparse
+import re
 import sys
-from typing import Sequence
+from typing import List, Sequence
 
 import pandas as pd
 from dateutil import tz
 
 from exabel_data_sdk import ExabelClient
 from exabel_data_sdk.client.api.data_classes.signal import Signal
 from exabel_data_sdk.scripts.csv_script import CsvScript
-from exabel_data_sdk.util.resource_name_normalization import to_entity_resource_names
+from exabel_data_sdk.util.resource_name_normalization import (
+    to_entity_resource_names,
+    validate_signal_name,
+)
 
 
 class LoadTimeSeriesFromCsv(CsvScript):
@@ -74,6 +78,34 @@ def run_script(self, client: ExabelClient, args: argparse.Namespace) -> None:
 
         print("Loading signals", ", ".join(str(s) for s in signals), "...")
 
+        # validate signal names
+        missing_header_pattern = re.compile(r"^Unnamed: ([0-9]+)$")
+        missing_headers: List[str] = []
+        invalid_signals: List[str] = []
+        for signal in signals:
+            try:
+                validate_signal_name(signal)
+            except ValueError:
+                # Pandas eats up any blank column names and replaces them with "Unnamed: N". Since
+                # this is invalid but not the actual column name, we give the end user a more
+                # precise error message
+                missing_header_match = missing_header_pattern.match(signal)
+                if missing_header_match:
+                    missing_headers.append(missing_header_match.group(1))
+                else:
+                    invalid_signals.append(signal)
+        if invalid_signals or missing_headers:
+            print(
+                "Encountered invalid signal names. Signal names must start with a letter, "
+                "and can only consist of letters, numbers, and underscore (_), and be "
+                "at most 64 characters"
+            )
+            if invalid_signals:
+                print(f"Invalid signal names: {', '.join(invalid_signals)}")
+            if missing_headers:
+                print(f"The following column(s) are missing headers: {', '.join(missing_headers)}")
+            sys.exit(1)
+
         prefix = "signals/"
         if args.namespace:
             prefix += args.namespace + "."

exabel_data_sdk/tests/scripts/test_load_time_series_from_csv.py

@@ -7,14 +7,9 @@
 
 from exabel_data_sdk import ExabelClient
 from exabel_data_sdk.scripts.load_time_series_from_csv import LoadTimeSeriesFromCsv
+from exabel_data_sdk.util.resource_name_normalization import validate_signal_name
 
-common_args = [
-    "script-name",
-    "--sep",
-    ";",
-    "--api-key",
-    "123",
-]
+common_args = ["script-name", "--sep", ";", "--api-key", "123"]
 
 
 class TestUploadTimeSeries(unittest.TestCase):
@@ -207,6 +202,35 @@ def test_read_file_with_integer_identifiers(self):
             check_freq=False,
         )
 
+    def test_should_fail_with_invalid_signal_names(self):
+        signals_errors = {
+            "0_starts_with_0": "Signal name must start with a letter, "
+            'contain only letters, numbers, and underscores, but got "0_starts_with_0"',
+            "contains_!llegal_chars": "Signal name must start with a letter, "
+            'contain only letters, numbers, and underscores, but got "contains_!llegal_chars"',
+            "": "Signal name cannot be empty",
+            "signal_with_sixty_five_characters_in_length_which_more_than_max__": "Signal name "
+            "cannot be longer than 64 characters, but got "
+            '"signal_with_sixty_five_characters_in_length_which_more_than_max__"',
+        }
+
+        for signal, error in signals_errors.items():
+            with self.assertRaises(ValueError) as cm:
+                validate_signal_name(signal)
+            self.assertEqual(str(cm.exception), error)
+
+    def test_valid_signal_names(self):
+        valid_signals = [
+            "signal",
+            "SIGNAL",
+            "signal_with_underscores",
+            "signal_1_with_underscores_and_numbers",
+            "signal_with_sixty_four_characters_in_length_which_is_the_maximum",
+        ]
+
+        for signal in valid_signals:
+            validate_signal_name(signal)
+
 
 if __name__ == "__main__":
     unittest.main()

exabel_data_sdk/util/resource_name_normalization.py

@@ -168,3 +168,19 @@ def to_entity_resource_names(
     result = identifiers.map(mapping)
     result.name = "entity"
     return result
+
+
+def validate_signal_name(name: str) -> None:
+    """
+    Validate that the given signal name is a legal signal name. A signal name is a string that
+    starts with a letter, and can contain letters, numbers, and underscores.
+    """
+    if not name:
+        raise ValueError("Signal name cannot be empty")
+    if len(name) > 64:
+        raise ValueError(f'Signal name cannot be longer than 64 characters, but got "{name}"')
+    if not re.match(r"^[a-zA-Z]\w{0,63}$", name):
+        raise ValueError(
+            f"Signal name must start with a letter, contain only letters, "
+            f'numbers, and underscores, but got "{name}"'
+        )