Add code for copying notes and images

.gitignore

@@ -9,3 +9,4 @@ tsconfig.tsbuildinfo
 private/
 .replit
 replit.nix
+misc/paths.py

misc/get_notes.py

@@ -0,0 +1,117 @@
+import mmap  # For checking files for the publish string without loading the whole thing into memory.
+import os
+import re
+import shutil  # For copying files.
+from paths import VAULT_PATH, REPO_PATH  # Private. VAULT_PATH just contains the path to my Obsidian vault root directory. REPO_PATH is the path to this repo.
+
+
+VAULT_EVERGREEN_PATH = os.fsencode(VAULT_PATH + "/Evergreen Thoughts")
+REPO_NOTES_PATH = os.fsencode(REPO_PATH + "/content")  # content is the Quartz directory I'll serve Evergreen Thoughts from. Nesting it further makes the URL weird.
+# TODO: Figure out how to make attachments (pics) work. I might need to nest that directory.
+
+def file_contains_tag(file_path, tag):
+    """Checks whether a file contains a tag.
+
+    set() is fine because Obsidian markdown files always fit into memory.
+    If that ever changes, seek an mmap implementation.
+    """
+
+    f = open(file_path, "r")
+    strings = set(f.read().split())  # Use split() not splitline() because Obsidian tags often have trailing whitespace
+    return tag in strings
+
+def get_image_filenames_from_md_file(file_path):
+    """Returns a (potentially empty) list of images linked from the file.
+
+    Example output list:
+    [('Screenshot_20221213-180612~2.png', ''), ('Screenshot_20221213-182124.png', '')]
+    """
+    output = []
+    f = open(file_path, "r")
+    f_read = f.read()
+
+    images_found = re.findall("!\[\[(.*?)\]\](.*)\s", f_read)  # This regex matches Obsidian-style markdown image links.
+    # It returns a (potentially empty) list of images linked from the file.
+    # Example output list:
+    # [('Screenshot_20221213-180612~2.png', ''), ('Screenshot_20221213-182124.png', '')]
+
+    # We clean the list here.
+    if images_found:
+        for image_tuple in images_found:
+            output.append(image_tuple[0])
+
+    return output
+
+
+
+
+files_scanned, count_publish, count_nopublish = 0, 0, 0
+
+for file_path in os.listdir(VAULT_EVERGREEN_PATH):
+    filename = os.fsdecode(file_path)
+    full_path = os.path.join(VAULT_EVERGREEN_PATH, file_path)
+
+    if filename.endswith(".md"):  # Filters out .DS_Store etc.
+        files_scanned += 1
+        if file_contains_tag(full_path, "#publish"):
+            count_publish += 1
+            print("PUBLISH", filename)
+            shutil.copy2(full_path, REPO_NOTES_PATH)
+        else:
+            count_nopublish += 1
+            print("SKIPPED", filename)
+
+
+print("===========")
+print("Finished!")
+print("- files_scanned:", files_scanned)
+print("- count_publish:", count_publish)
+print("- count_nopublish:", count_nopublish)
+
+# Next, do another pass for safety that deletes any files that have nopublish or maybepublish in them.
+files_scanned, count_left_alone, count_removed = 0, 0, 0
+
+for file_path in os.listdir(REPO_NOTES_PATH):
+    filename = os.fsdecode(file_path)
+    full_path = os.path.join(REPO_NOTES_PATH, file_path)
+
+    if filename.endswith(".md"):  # Filters out .DS_Store etc.
+        files_scanned += 1
+        if file_contains_tag(full_path, "#nopublish"):
+        	# TODO: Make this also do #maybepublish
+            count_removed += 1
+            print("REMOVED", filename)
+            os.remove(full_path)
+        else:
+            count_left_alone += 1
+            # Usually they are left alone so no need to write them all
+            # print("LEFT AL", filename)
+
+print("===========")
+print("Finished!")
+print("- files_scanned:", files_scanned)
+print("- count_removed:", count_removed)
+print("- count_left_alone:", count_left_alone)
+
+
+# now, time to get the images.
+
+
+
+######
+image_filenames = []  # this is just the string. It fits in memory
+files_copied = 0
+
+for file_path in os.listdir(REPO_NOTES_PATH):
+    filename = os.fsdecode(file_path)
+    full_path = os.path.join(REPO_NOTES_PATH, file_path)
+
+    if filename.endswith(".md"):  # Filters out .DS_Store etc.
+        image_filenames += get_image_filenames_from_md_file(full_path)
+
+for image_filename in image_filenames:
+    image_path = os.fsencode(VAULT_PATH + "/Attachments/" + image_filename)
+    print(image_path)
+    shutil.copy2(image_path, REPO_NOTES_PATH)
+
+

misc/paths.py

@@ -0,0 +1,2 @@
+VAULT_PATH='/Users/warren/PathToYourObsidianVault'
+REPO_PATH='/Users/warren/PathTo/perdicis-notes'