Fix generic_json parser (#1365)

Fixes #1347.

archivebox/parsers/generic_json.py

@@ -18,9 +18,16 @@ def parse_generic_json_export(json_file: IO[str], **_kwargs) -> Iterable[Link]:
 
     json_file.seek(0)
 
-    # sometimes the first line is a comment or filepath, so we get everything after the first {
-    json_file_json_str = '{' + json_file.read().split('{', 1)[-1]
-    links = json.loads(json_file_json_str)
+    try:
+        links = json.load(json_file)
+    except json.decoder.JSONDecodeError:
+        # sometimes the first line is a comment or other junk, so try without
+        json_file.seek(0)
+        first_line = json_file.readline()
+        #print('      > Trying JSON parser without first line: "', first_line.strip(), '"', sep= '')
+        links = json.load(json_file)
+        # we may fail again, which means we really don't know what to do
+
     json_date = lambda s: datetime.strptime(s, '%Y-%m-%dT%H:%M:%S%z')
 
     for link in links:
@@ -59,11 +66,20 @@ def parse_generic_json_export(json_file: IO[str], **_kwargs) -> Iterable[Link]:
             elif link.get('name'):
                 title = link['name'].strip()
 
+            # if we have a list, join it with commas
+            tags = link.get('tags')
+            if type(tags) == list:
+                tags = ','.join(tags)
+            elif type(tags) == str:
+                # if there's no comma, assume it was space-separated
+                if ',' not in tags:
+                    tags = tags.replace(' ', ',')
+
             yield Link(
                 url=htmldecode(url),
                 timestamp=ts_str,
                 title=htmldecode(title) or None,
-                tags=htmldecode(link.get('tags')) or '',
+                tags=htmldecode(tags),
                 sources=[json_file.name],
             )
 

tests/mock_server/templates/example.json

@@ -0,0 +1,6 @@
+[
+{"href":"http://127.0.0.1:8080/static/example.com.html","description":"Example","extended":"","meta":"18a973f09c9cc0608c116967b64e0419","hash":"910293f019c2f4bb1a749fb937ba58e3","time":"2014-06-14T15:51:42Z","shared":"no","toread":"no","tags":"Tag1 Tag2","trap":"http://www.example.com/should-not-exist"},
+{"href":"http://127.0.0.1:8080/static/iana.org.html","description":"Example 2","extended":"","meta":"18a973f09c9cc0608c116967b64e0419","hash":"910293f019c2f4bb1a749fb937ba58e3","time":"2014-06-14T15:51:43Z","shared":"no","toread":"no","tags":"Tag3,Tag4 with Space"},
+{"href":"http://127.0.0.1:8080/static/shift_jis.html","description":"Example 2","extended":"","meta":"18a973f09c9cc0608c116967b64e0419","hash":"910293f019c2f4bb1a749fb937ba58e3","time":"2014-06-14T15:51:44Z","shared":"no","toread":"no","tags":["Tag5","Tag6 with Space"]},
+{"href":"http://127.0.0.1:8080/static/title_og_with_html","description":"Example 2","extended":"","meta":"18a973f09c9cc0608c116967b64e0419","hash":"910293f019c2f4bb1a749fb937ba58e3","time":"2014-06-14T15:51:45Z","shared":"no","toread":"no"}
+]

tests/mock_server/templates/example.json.bad

@@ -0,0 +1,2 @@
+this line would cause problems but --parser=json will actually skip it
+[{"href":"http://127.0.0.1:8080/static/example.com.html","description":"Example","extended":"","meta":"18a973f09c9cc0608c116967b64e0419","hash":"910293f019c2f4bb1a749fb937ba58e3","time":"2014-06-14T15:51:42Z","shared":"no","toread":"no","tags":"Tag1 Tag2","trap":"http://www.example.com/should-not-exist"}]

tests/test_add.py

@@ -91,3 +91,60 @@ def test_extract_input_uses_only_passed_extractors(tmp_path, process):
 
     assert (archived_item_path / "warc").exists()
     assert not (archived_item_path / "singlefile.html").exists()
+
+def test_json(tmp_path, process, disable_extractors_dict):
+    with open('../../mock_server/templates/example.json', 'r', encoding='utf-8') as f:
+        arg_process = subprocess.run(
+            ["archivebox", "add", "--index-only", "--parser=json"],
+            stdin=f,
+            capture_output=True,
+            env=disable_extractors_dict,
+        )
+
+    conn = sqlite3.connect("index.sqlite3")
+    c = conn.cursor()
+    urls = c.execute("SELECT url from core_snapshot").fetchall()
+    tags = c.execute("SELECT name from core_tag").fetchall()
+    conn.commit()
+    conn.close()
+
+    urls = list(map(lambda x: x[0], urls))
+    assert "http://127.0.0.1:8080/static/example.com.html" in urls
+    assert "http://127.0.0.1:8080/static/iana.org.html" in urls
+    assert "http://127.0.0.1:8080/static/shift_jis.html" in urls
+    assert "http://127.0.0.1:8080/static/title_og_with_html" in urls
+    # if the following URL appears, we must have fallen back to another parser
+    assert not "http://www.example.com/should-not-exist" in urls
+
+    tags = list(map(lambda x: x[0], tags))
+    assert "Tag1" in tags
+    assert "Tag2" in tags
+    assert "Tag3" in tags
+    assert "Tag4 with Space" in tags
+    assert "Tag5" in tags
+    assert "Tag6 with Space" in tags
+
+def test_json_with_leading_garbage(tmp_path, process, disable_extractors_dict):
+    with open('../../mock_server/templates/example.json.bad', 'r', encoding='utf-8') as f:
+        arg_process = subprocess.run(
+            ["archivebox", "add", "--index-only", "--parser=json"],
+            stdin=f,
+            capture_output=True,
+            env=disable_extractors_dict,
+        )
+
+    conn = sqlite3.connect("index.sqlite3")
+    c = conn.cursor()
+    urls = c.execute("SELECT url from core_snapshot").fetchall()
+    tags = c.execute("SELECT name from core_tag").fetchall()
+    conn.commit()
+    conn.close()
+
+    urls = list(map(lambda x: x[0], urls))
+    assert "http://127.0.0.1:8080/static/example.com.html" in urls
+    # if the following URL appears, we must have fallen back to another parser
+    assert not "http://www.example.com/should-not-exist" in urls
+
+    tags = list(map(lambda x: x[0], tags))
+    assert "Tag1" in tags
+    assert "Tag2" in tags