From 6e844a3cb185d1a433d458a88ca7e51b2a5cd3b8 Mon Sep 17 00:00:00 2001
From: tiyn <tiyn@mail-mk.eu>
Date: Mon, 22 Apr 2024 02:06:12 +0200
Subject: [PATCH] added tts functionality

---
 Dockerfile               | 18 ++++++++------
 README.md                |  1 +
 src/app.py               |  1 +
 src/content.py           | 54 ++++++++++++++++++++++++++++++++++++++++
 src/requirements.txt     |  1 +
 src/static/css/style.css |  6 +++++
 src/static/tmp/.gitkeep  |  0
 7 files changed, 73 insertions(+), 8 deletions(-)
 create mode 100644 src/static/tmp/.gitkeep

diff --git a/Dockerfile b/Dockerfile
index 1dd726f..37d0de1 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -2,11 +2,9 @@ FROM python:3
 
 MAINTAINER tiyn tiyn@mail-mk.eu
 
-COPY src /blog
-
-WORKDIR /blog
-
-RUN pip3 install -r requirements.txt
+ENV LANG en_US.UTF-8
+ENV LANGUAGE en_US:en
+ENV LC_ALL en_US.UTF-8
 
 RUN apt-get update && \
     apt-get install -y locales && \
@@ -14,9 +12,13 @@ RUN apt-get update && \
     sed -i '/en_US.UTF-8/s/^# //g' /etc/locale.gen && \
     dpkg-reconfigure --frontend=noninteractive locales
 
-ENV LANG en_US.UTF-8
-ENV LANGUAGE en_US:en
-ENV LC_ALL en_US.UTF-8
+RUN apt-get install -y espeak
+
+COPY src /blog
+
+WORKDIR /blog
+
+RUN pip3 install -r requirements.txt
 
 VOLUME /blog/templates/entry
 
diff --git a/README.md b/README.md
index 5ccc114..9e0b362 100644
--- a/README.md
+++ b/README.md
@@ -22,6 +22,7 @@ via plain text files.
   - [x] Links to standalone article
 - [x] Standalone article page
   - [x] Links to scrolling blog page
+  - [x] TTS Functionality
 - [x] RSS feed
 - [x] Navigation
   - [x] Header
diff --git a/src/app.py b/src/app.py
index 8035d32..60c1495 100644
--- a/src/app.py
+++ b/src/app.py
@@ -137,4 +137,5 @@ def feed():
 
 
 if __name__ == "__main__":
+  con_gen.prepare_tts()
   app.run(host="0.0.0.0")
diff --git a/src/content.py b/src/content.py
index bc92506..5ec7cb5 100644
--- a/src/content.py
+++ b/src/content.py
@@ -1,3 +1,4 @@
+import glob
 import locale
 import os
 import pathlib
@@ -7,6 +8,7 @@ from os import path
 
 import markdown
 from bs4 import BeautifulSoup
+from gtts import gTTS
 
 import config
 import search
@@ -144,6 +146,11 @@ def gen_stand_string(path_ex):
     content_string += "<a href=\"" + "/index.html#" + \
         filename_no_end + "\">" + curr_date + "</a>"
     content_string += "<br><br>\n"
+    if os.path.isfile("static/tmp/" + filename_no_end + ".mp3"):
+      content_string += "<audio controls>\n"
+      content_string += '<source src="/static/tmp/' + filename_no_end + '.mp3" type="audio/mp3">\n'
+      content_string += "</audio>\n"
+    content_string += "<br><br>\n"
     if filename.endswith(".html"):
       for line in text:
         content_string += line
@@ -270,3 +277,50 @@ def create_preview(path, is_markdown):
     preview = "\n<p>" + first_p.text + "</p>\n"
   preview += "...<br>"
   return preview
+
+
+def get_text_only(filename):
+  """
+  Convert a file to text only to use in tts
+
+  Parameters:
+  path (string): path to the article
+
+  Returns:
+  string: unformatted string containing the contents of the file
+  """
+  # filename = os.path.join(ENTRY_DIR, path)
+  clean_text = ""
+  if path.exists(filename):
+    title = open(filename).readline().rstrip("\n")
+    text = open(filename).readlines()[1:]
+    filename_no_end = filename.split(".", 1)[0]
+    filename_no_end = filename_no_end.split("/")[-1]
+    content_string = ""
+    if filename.endswith(".html"):
+      for line in text:
+        content_string += line
+    if filename.endswith(".md"):
+      content_string += gen_md_content(filename, 1)
+    content_string = absolutize_html(content_string)
+    soup = BeautifulSoup(content_string, "html.parser")
+    tag_to_remove = soup.find("figure")
+    if tag_to_remove:
+      tag_to_remove.decompose()
+    clean_text = soup.get_text(separator=" ")
+    clean_text = title + "\n\n" + clean_text
+  return clean_text
+
+
+def prepare_tts():
+  files = glob.glob('static/tmp/*')
+  for f in files:
+    os.remove(f)
+  files = glob.glob('templates/entry/*')
+  clean_text = ""
+  for f in files:
+    clean_text = get_text_only(f)
+    tts = gTTS(clean_text, lang=LANGUAGE.split("-")[0])
+    _, tail = os.path.split(f)
+    new_filename = "static/tmp/" + os.path.splitext(tail)[0] + ".mp3"
+    tts.save(new_filename)
diff --git a/src/requirements.txt b/src/requirements.txt
index 13dc9a4..71917a2 100644
--- a/src/requirements.txt
+++ b/src/requirements.txt
@@ -5,3 +5,4 @@ WTForms
 Flask_WTF
 Font-Awesome-Flask
 BeautifulSoup4
+gTTS
diff --git a/src/static/css/style.css b/src/static/css/style.css
index 6cbfb52..65d5d8b 100644
--- a/src/static/css/style.css
+++ b/src/static/css/style.css
@@ -269,6 +269,12 @@ form.search::after {
 .standalone img {
   width: 50%;
 }
+
 .entry img {
   width: 50%;
 }
+
+audio {
+    width: 50%;
+    height: 5vh;
+}
diff --git a/src/static/tmp/.gitkeep b/src/static/tmp/.gitkeep
new file mode 100644
index 0000000..e69de29