py2ls 0.1.7.0__tar.gz → 0.1.7.2__tar.gz

{py2ls-0.1.7.0 → py2ls-0.1.7.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: py2ls
-Version: 0.1.7.0
+Version: 0.1.7.2
 Summary: py(thon)2(too)ls
 Author: Jianfeng
 Author-email: Jianfeng.Liu0413@gmail.com

{py2ls-0.1.7.0 → py2ls-0.1.7.2}/py2ls/netfinder.py

@@ -115,7 +115,7 @@ def extract_text_from_content(content, content_type="text/html", where=None, wha
             if extend:
                 texts = ""
                 for tag in result_set:
-                    texts += extract_text(tag) + "\n"
+                    texts =texts+" "+ extract_text(tag) + "\n"
                 text_list = [tx.strip() for tx in texts.split("\n") if tx.strip()]
                 return text_list
             else:
@@ -1062,37 +1062,33 @@ def find_all(url, dir_save=None, driver='req'):
     return df
 
 
-def flist(fpath, filter="all"):
+def flist(fpath, kind="all"):
     all_files = [os.path.join(fpath, f) for f in os.listdir(fpath) if os.path.isfile(os.path.join(fpath, f))]
-    if isinstance(filter, str):
-        filter=[filter]
-    if isinstance(filter, list):
-        filt_files=[]
-        for filter_ in filter:
-            filt_files.extend(flist(fpath, filter_))
-        return filt_files
-    else:
-        if 'all' in filter.lower():
-            return all_files
-        else:
-            filt_files = [f for f in all_files if istype(f, filter)]
-        return filt_files
+    if filter == "all" or 'all' in filter:
+        return all_files
+    if isinstance(kind, str):
+        kind=[kind] 
+    filt_files=[]
+    for kind_ in kind:
+        if isa(f, kind_):
+            filt_files.extend(flist(fpath, kind_))
+    return filt_files
 
-def istype(fpath, filter='img'):
+def isa(fpath, kind='img'):
     """
-    Filters file paths based on the specified filter.
+    kinds file paths based on the specified kind.
     Args:
         fpath (str): Path to the file.
-        filter (str): filter of file to filter. Default is 'img' for images. Other options include 'doc' for documents,
+        kind (str): kind of file to kind. Default is 'img' for images. Other options include 'doc' for documents,
                     'zip' for ZIP archives, and 'other' for other types of files.
     Returns:
-        bool: True if the file matches the filter, False otherwise.
+        bool: True if the file matches the kind, False otherwise.
     """
-    if 'img' in filter.lower():
+    if 'img' in kind.lower():
         return is_image(fpath)
-    elif 'doc' in filter.lower():
+    elif 'doc' in kind.lower():
         return is_document(fpath)
-    elif 'zip' in filter.lower():
+    elif 'zip' in kind.lower():
         return is_zip(fpath)
     else:
         return False

{py2ls-0.1.7.0 → py2ls-0.1.7.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "py2ls"
-version = "0.1.7.0"
+version = "0.1.7.2"
 description = "py(thon)2(too)ls"
 authors = ["Jianfeng <Jianfeng.Liu0413@gmail.com>"]
 readme = "README.md"