fix behavior and tests

tests/unit_tests.py

@@ -73,11 +73,12 @@ def load_mock_page(url, xml_flag=False, langcheck=None, tei_output=False):
                 htmlstring = htmlbinary
         else:
             print('Encoding error')
-    output_format = 'txt'
-    if xml_flag is True:
+    if xml_flag:
         output_format = 'xml'
-    if tei_output is True:
-        output_format = 'tei'
+    elif tei_output:
+        output_format = 'xmltei'
+    else:
+        output_format = 'txt'
     return extract(htmlstring, url,
                      record_id='0000',
                      no_fallback=False,

trafilatura/cli.py

@@ -17,7 +17,7 @@
                         file_processing_pipeline, load_blacklist,
                         load_input_dict, probe_homepage,
                         url_processing_pipeline, write_result)
-from .settings import PARALLEL_CORES, SUPPORTED_FORMATS
+from .settings import PARALLEL_CORES, SUPPORTED_FMT_CLI
 
 # fix output encoding on some systems
 try:
@@ -162,10 +162,10 @@ def add_args(parser):
     # https://docs.python.org/3/library/argparse.html#argparse.ArgumentParser.add_mutually_exclusive_group
     group5_ex.add_argument("-out",
                         help=argparse.SUPPRESS,
-                        choices=sorted(SUPPORTED_FORMATS))
+                        choices=sorted(SUPPORTED_FMT_CLI))
     group5_ex.add_argument('--output-format',
                         help="determine output format",
-                        choices=sorted(SUPPORTED_FORMATS),
+                        choices=sorted(SUPPORTED_FMT_CLI),
                         default='txt')
     group5_ex.add_argument("--csv",
                         help="shorthand for CSV output",

trafilatura/core.py

@@ -25,6 +25,8 @@
 
 LOGGER = logging.getLogger(__name__)
 
+TXT_FORMATS = {"markdown", "txt"}
+
 
 def determine_returnstring(document, options):
     '''Convert XML tree to chosen format, clean the result and output it as a string'''
@@ -347,7 +349,7 @@ def extract(filecontent, url=None, record_id=None, no_fallback=False,
     if document is None:
         return None
 
-    if options.format not in ("markdown", "txt"):
+    if options.format not in TXT_FORMATS:
         # add record ID to metadata
         document.id = record_id
         # calculate fingerprint

trafilatura/settings.py

@@ -22,7 +22,8 @@
 from .utils import line_processing
 
 
-SUPPORTED_FORMATS = {"csv", "json", "html", "markdown", "txt", "xml", "xmltei"}
+SUPPORTED_FMT_CLI = ["csv", "json", "html", "markdown", "txt", "xml", "xmltei"]
+SUPPORTED_FORMATS = set(SUPPORTED_FMT_CLI) | {"python"}  # the latter for bare_extraction() only
 _SUPPORTED = ', '.join(sorted(SUPPORTED_FORMATS))