Fix/handle-spooled-temp-file-eml (#800)

This PR is for the unstructured-api smoke tests pass.

CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.7.8-dev1
+## 0.7.8
 
 ### Enhancements
 
@@ -8,6 +8,7 @@
 
 ### Fixes
 
+* Updates the `parse_email` for `partition_eml` so that `unstructured-api` passes the smoke tests
 * `partition_email` now works if there is no message content
 * Updates the `"fast"` strategy for `partition_pdf` so that it's able to recursively
 * Adds recursive functionality to all fsspec connectors

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.7.8-dev1"  # pragma: no cover
+__version__ = "0.7.8"  # pragma: no cover

unstructured/file_utils/encoding.py

@@ -2,6 +2,8 @@ from typing import IO, Optional, Tuple, Union
 
 import chardet
 
+from unstructured.partition.common import convert_to_bytes
+
 ENCODE_REC_THRESHOLD = 0.5
 
 # popular encodings from https://en.wikipedia.org/wiki/Popularity_of_text_encodings
@@ -46,14 +48,7 @@ def detect_file_encoding(
         with open(filename, "rb") as f:
             byte_data = f.read()
     elif file:
-        if isinstance(file, bytes):
-            byte_data = file
-        else:
-            if not hasattr(file, "mode") or "b" in file.mode:
-                byte_data = file.read()
-            else:
-                with open(file.name, "rb") as f:
-                    byte_data = f.read()
+        byte_data = convert_to_bytes(file)
     else:
         raise FileNotFoundError("No filename nor file were specified")
 

unstructured/partition/common.py

@@ -1,9 +1,9 @@
 from __future__ import annotations
 
 import subprocess
-from io import BytesIO
+from io import BufferedReader, BytesIO, TextIOWrapper
 from tempfile import SpooledTemporaryFile
-from typing import TYPE_CHECKING, Any, BinaryIO, Dict, List, Optional, Tuple, Union
+from typing import IO, TYPE_CHECKING, Any, BinaryIO, Dict, List, Optional, Tuple, Union
 
 from docx import table as docxtable
 from tabulate import tabulate
@@ -184,6 +184,25 @@ def spooled_to_bytes_io_if_needed(
         return file_obj
 
 
+def convert_to_bytes(
+    file: Optional[Union[bytes, SpooledTemporaryFile, IO]] = None,
+) -> bytes:
+    if isinstance(file, bytes):
+        f_bytes = file
+    elif isinstance(file, SpooledTemporaryFile):
+        file.seek(0)
+        f_bytes = file.read()
+    elif isinstance(file, BytesIO):
+        f_bytes = file.getvalue()
+    elif isinstance(file, (TextIOWrapper, BufferedReader)):
+        with open(file.name, "rb") as f:
+            f_bytes = f.read()
+    else:
+        raise ValueError("Invalid file-like object type")
+
+    return f_bytes
+
+
 def convert_ms_office_table_to_text(table: docxtable.Table, as_html: bool = True):
     """
     Convert a table object from a Word document to an HTML table string using the tabulate library.

unstructured/partition/email.py

@@ -4,6 +4,7 @@ import re
 import sys
 from email.message import Message
 from functools import partial
+from tempfile import SpooledTemporaryFile
 from typing import IO, Dict, List, Optional, Tuple, Union
 
 from unstructured.file_utils.encoding import (
@@ -11,7 +12,10 @@ from unstructured.file_utils.encoding import (
     format_encoding_str,
     read_txt_file,
 )
-from unstructured.partition.common import exactly_one
+from unstructured.partition.common import (
+    convert_to_bytes,
+    exactly_one,
+)
 
 if sys.version_info < (3, 8):
     from typing_extensions import Final
@@ -189,14 +193,14 @@ def find_embedded_image(
 
 def parse_email(
     filename: Optional[str] = None,
-    file: Optional[IO] = None,
+    file: Optional[Union[IO, SpooledTemporaryFile]] = None,
 ) -> Tuple[Optional[str], Message]:
     if filename is not None:
         with open(filename, "rb") as f:
             msg = email.message_from_binary_file(f)
     elif file is not None:
-        with open(file.name, "rb") as f:
-            msg = email.message_from_binary_file(f)
+        f_bytes = convert_to_bytes(file)
+        msg = email.message_from_bytes(f_bytes)
     else:
         raise ValueError("Either 'filename' or 'file' must be provided.")
 
@@ -216,7 +220,7 @@ def parse_email(
 @add_metadata_with_filetype(FileType.EML)
 def partition_email(
     filename: Optional[str] = None,
-    file: Optional[IO] = None,
+    file: Optional[Union[IO, SpooledTemporaryFile]] = None,
     text: Optional[str] = None,
     content_source: str = "text/html",
     encoding: Optional[str] = None,