Not everyone has Javascript turned on, so I wanted
to start with something that made sense even in
RSS feeds like the one on Planet Emacslife (which
strips out <style>
and <script>
) and was
progressively enhanced with captions and
highlighting if you saw it on my site.
The blog aggregator I'm using, Planet Venus, hasn't been updated in 14 years. It even uses Python 2. I considered switching to a different aggregator, so I started checking out different community planets. Most of the other planets listed in this HN thread about aggregators looked like they were using the same Planet Venus aggregator, although these were some planets that used something else:
I decided I'd stick with Planet Venus for now, since I could probably figure out how to get the attributes sorted out.
I found planet/vendor/feedparser.py
by digging
around. Adding mix-blend-mode
to the list of
attributes there was not enough to get it working.
I started exploring pdb for interactive Python
debugging inside Emacs, although I think dap is an
option too. I wrote a short bit of code to test things out:
import sys,os sys.path.insert(1, os.path.join(os.path.dirname(__file__), 'planet/vendor')) from feedparser import _sanitizeHTML assert 'strong' in _sanitizeHTML('<strong>Hello</strong>', 'utf-8', 'text/html') assert 'mix-blend-mode' in _sanitizeHTML('<svg><path style="fill: red;mix-blend-mode:darken"></path></svg>', 'utf-8', 'text/html')
It was pretty easy to use pdb to start stepping
through and into functions, although I didn't dig
into it deeply because I figured it out another
way. While looking through the pull requests for
the Venus repository, I came across this
pull
request to add data- attributes which was
helpful because it pointed me to
planet/vendor/html5lib/sanitizer.py
. Once I
added mix-blend-mode
to that one, things worked.
Here's my Github branch.
On a somewhat related note, I also had to patching shr to handle SVG images with viewBox attributes. I guess SVGs aren't that common yet, but I'm looking forward to playing around with them more, so I might as well make things better (at least when it comes to things I can actually tweak). mix-blend-mode on SVG elements says it's not supported in Safari or a bunch of mobile browsers, but it seems to be working on my phone, so maybe that's cool now. Using mix-blend-mode means I don't have to do something complicated when it comes to animating highlights while still keeping text visible, and improving SVG support is the right thing to do. Onward!
]]>import pandas as pd import matplotlib.pyplot as plt df = pd.DataFrame(data[1:], columns=data[0]) df = df.drop('Weeks to CFP', axis=1).groupby(['Year']).sum() fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(12,6)) fig1 = df['Count'].plot(kind="bar", ax=ax[0], title='Number of submissions') fig2 = df['Minutes'].plot(kind="bar", ax=ax[1], title='Number of minutes') fig.get_figure().savefig('emacsconf-by-year.png') return df
Year | Count | Minutes |
---|---|---|
2019 | 28 | 429 |
2020 | 35 | 699 |
2021 | 44 | 578 |
2022 | 29 | 512 |
2023 | 39 | 730 |
I also wanted to make an animated GIF so that the cumulative graphs could be a little easier to understand.
import pandas as pd import matplotlib.pyplot as plt import imageio as io df = pd.DataFrame(data[1:], columns=data[0]) fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(12,6)) count = pd.pivot_table(df, columns=['Year'], index=['Weeks to CFP'], values='Count', aggfunc='sum', fill_value=0).iloc[::-1].sort_index(ascending=True).cumsum() minutes = pd.pivot_table(df, columns=['Year'], index=['Weeks to CFP'], values='Minutes', aggfunc='sum', fill_value=0).iloc[::-1].sort_index(ascending=True).cumsum() ax[0].set_ylim([0, count.max().max()]) ax[1].set_ylim([0, minutes.max().max()]) with io.get_writer('emacsconf-combined.gif', mode='I', duration=[500, 500, 500, 500, 1000], loop=0) as writer: for year in range(2019, 2024): count[year].plot(ax=ax[0], title='Cumulative submissions') minutes[year].plot(ax=ax[1], title='Cumulative minutes') ax[0].legend(loc='upper left') ax[1].legend(loc='upper left') for axis in ax: for line in axis.get_lines(): if line.get_label() == '2023': line.set_linewidth(5) for line in axis.legend().get_lines(): if line.get_label() == '2023': line.set_linewidth(5) filename = f'emacsconf-combined-${year}.png' fig.get_figure().savefig(filename) image = io.v3.imread(filename) writer.append_data(image)
I am not quite sure what kind of story this data tells (aside from the fact that there sure are a lot of great talks), but it was fun to learn how to make more kinds of graphs and animate them too. Could be useful someday. =)
]]>
This year, I managed to not panic and I also resisted the urge to extend the
CFP deadline, trusting that there will actually be tons of cool stuff.
It helped that my schedule SVG code let me visualize what the
conference could feel like with the submissions so far, so we started
with a reasonably nice one-track conference and built up from there.
It also helped that I'd gone back to the submissions for 2022 and
plotted them by the number of weeks before the CFP deadline, and I
knew that there'd be a big spike from all those people whose Org
DEADLINE:
properties would nudge them into finalizing their
proposals.
Out of curiosity, I wanted to see how the stats for this year compared with previous years. I wrote a small function to collect the data that I wanted to summarize:
(defun emacsconf-count-submissions-by-week (&optional info cfp-deadline) "Count submissions in INFO by distance to CFP-DEADLINE." (setq cfp-deadline (or cfp-deadline emacsconf-cfp-deadline)) (setq info (or info (emacsconf-get-talk-info))) (cons '("Weeks to CFP end date" "Count" "Hours") (mapcar (lambda (entry) (list (car entry) (length (cdr entry)) (apply '+ (mapcar 'cdr (cdr entry))))) (seq-group-by 'car (sort (seq-keep (lambda (o) (and (emacsconf-publish-talk-p o) (plist-get o :date-submitted) (cons (floor (/ (days-between (plist-get o :date-submitted) cfp-deadline) 7.0)) (string-to-number (or (plist-get o :video-duration) (plist-get o :time) "0"))))) info) (lambda (a b) (< (car a) (car b))))))))
and then I ran it against the different files for each year, filling in the previous years' data as needed. The resulting table is pretty long, so I've put that in a collapsible section.
(let ((years `((2023 "~/proj/emacsconf/2023/private/conf.org" "2023-09-15") (2022 "~/proj/emacsconf/2022/private/conf.org" "2022-09-18") (2021 "~/proj/emacsconf/2021/private/conf.org" "2021-09-30") (2020 "~/proj/emacsconf/wiki/2020/submissions.org" "2020-09-30") (2019 "~/proj/emacsconf/2019/private/conf.org" "2019-08-31")))) (append '(("Weeks to CFP" "Year" "Count" "Minutes")) (seq-mapcat (lambda (year-info) (let ((emacsconf-org-file (elt year-info 1)) (emacsconf-cfp-deadline (elt year-info 2)) (year (car year-info))) (mapcar (lambda (o) (list (car o) year (cadr o) (elt o 2))) (cdr (emacsconf-count-submissions-by-week (emacsconf-get-talk-info) emacsconf-cfp-deadline))))) years)))
Weeks to CFP | Year | Count | Minutes |
-12 | 2023 | 4 | 70 |
-9 | 2023 | 2 | 30 |
-7 | 2023 | 2 | 30 |
-5 | 2023 | 2 | 30 |
-4 | 2023 | 2 | 60 |
-3 | 2023 | 3 | 40 |
-2 | 2023 | 5 | 130 |
-1 | 2023 | 10 | 180 |
0 | 2023 | 8 | 140 |
1 | 2023 | 1 | 20 |
-8 | 2022 | 2 | 25 |
-5 | 2022 | 2 | 31 |
-3 | 2022 | 2 | 31 |
-2 | 2022 | 2 | 17 |
-1 | 2022 | 8 | 191 |
0 | 2022 | 8 | 110 |
1 | 2022 | 5 | 107 |
-8 | 2021 | 4 | 50 |
-7 | 2021 | 2 | 17 |
-6 | 2021 | 1 | 7 |
-5 | 2021 | 2 | 22 |
-4 | 2021 | 2 | 19 |
-3 | 2021 | 5 | 73 |
-2 | 2021 | 1 | 10 |
-1 | 2021 | 12 | 163 |
0 | 2021 | 13 | 197 |
1 | 2021 | 1 | 10 |
2 | 2021 | 1 | 10 |
-5 | 2020 | 1 | 10 |
-4 | 2020 | 1 | 15 |
-2 | 2020 | 1 | 30 |
-1 | 2020 | 4 | 68 |
0 | 2020 | 21 | 424 |
1 | 2020 | 7 | 152 |
-5 | 2019 | 2 | 45 |
-4 | 2019 | 1 | 21 |
-2 | 2019 | 6 | 126 |
-1 | 2019 | 9 | 82 |
0 | 2019 | 9 | 148 |
2 | 2019 | 1 | 7 |
Some talks were proposed off-list and are not captured here, and cancelled or withdrawn talks weren't included either. The times for previous years use the actual video time, and the times for this year use proposed times.
Off the top of my head, I didn't know of an easy way to make a pivot
table or cross-tab using just Org Mode or Emacs Lisp. I tried using
datamash, but I was having a hard time getting my output just the way
I wanted it. Fortunately, it was super-easy to get my data from an Org
table into Python so I could use pandas.pivot_table. Because I had
used #+NAME: submissions-by-week
to label the table, I could use
:var data=submissions-by-week
to refer to the data in my Python
program. Then I could summarize them by week.
Here's the number of submissions by the number of weeks to the original CFP deadline, so we can see people generally like to target the CFP date.
import pandas as pd import matplotlib.pyplot as plt df = pd.DataFrame(data[1:], columns=data[0]) df = pd.pivot_table(df, columns=['Year'], index=['Weeks to CFP'], values='Count', aggfunc='sum', fill_value=0).iloc[::-1].sort_index(ascending=True) fig, ax = plt.subplots() figure = df.plot(title='Number of submissions by number of weeks to the CFP end date', ax=ax) for line in ax.get_lines(): if line.get_label() == '2023': line.set_linewidth(5) for line in plt.legend().get_lines(): if line.get_label() == '2023': line.set_linewidth(5) figure.get_figure().savefig('number-of-submissions.png') return df
Weeks to CFP | 2019 | 2020 | 2021 | 2022 | 2023 |
---|---|---|---|---|---|
-12 | 0 | 0 | 0 | 0 | 4 |
-9 | 0 | 0 | 0 | 0 | 2 |
-8 | 0 | 0 | 4 | 2 | 0 |
-7 | 0 | 0 | 2 | 0 | 2 |
-6 | 0 | 0 | 1 | 0 | 0 |
-5 | 2 | 1 | 2 | 2 | 2 |
-4 | 1 | 1 | 2 | 0 | 2 |
-3 | 0 | 0 | 5 | 2 | 3 |
-2 | 6 | 1 | 1 | 2 | 5 |
-1 | 9 | 4 | 12 | 8 | 10 |
0 | 9 | 21 | 13 | 8 | 8 |
1 | 0 | 7 | 1 | 5 | 1 |
2 | 1 | 0 | 1 | 0 | 0 |
Calculating the cumulative number of submissions might be more useful. Here, each row shows the number received so far.
import pandas as pd import matplotlib.pyplot as plt df = pd.DataFrame(data[1:], columns=data[0]) df = pd.pivot_table(df, columns=['Year'], index=['Weeks to CFP'], values='Count', aggfunc='sum', fill_value=0).iloc[::-1].sort_index(ascending=True).cumsum() fig, ax = plt.subplots() figure = df.plot(title='Cumulative submissions by number of weeks to the CFP end date', ax=ax) for line in ax.get_lines(): if line.get_label() == '2023': line.set_linewidth(5) for line in plt.legend().get_lines(): if line.get_label() == '2023': line.set_linewidth(5) figure.get_figure().savefig('cumulative-submissions.png') return df
Weeks to CFP | 2019 | 2020 | 2021 | 2022 | 2023 |
---|---|---|---|---|---|
-12 | 0 | 0 | 0 | 0 | 4 |
-9 | 0 | 0 | 0 | 0 | 6 |
-8 | 0 | 0 | 4 | 2 | 6 |
-7 | 0 | 0 | 6 | 2 | 8 |
-6 | 0 | 0 | 7 | 2 | 8 |
-5 | 2 | 1 | 9 | 4 | 10 |
-4 | 3 | 2 | 11 | 4 | 12 |
-3 | 3 | 2 | 16 | 6 | 15 |
-2 | 9 | 3 | 17 | 8 | 20 |
-1 | 18 | 7 | 29 | 16 | 30 |
0 | 27 | 28 | 42 | 24 | 38 |
1 | 27 | 35 | 43 | 29 | 39 |
2 | 28 | 35 | 44 | 29 | 39 |
And here's the cumulative number of minutes based on the proposals.
import pandas as pd import matplotlib.pyplot as plt df = pd.DataFrame(data[1:], columns=data[0]) df = pd.pivot_table(df, columns=['Year'], index=['Weeks to CFP'], values='Minutes', aggfunc='sum', fill_value=0).iloc[::-1].sort_index(ascending=True).cumsum() fig, ax = plt.subplots() figure = df.plot(title='Cumulative minutes by number of weeks to the CFP end date', ax=ax) for line in ax.get_lines(): if line.get_label() == '2023': line.set_linewidth(5) for line in plt.legend().get_lines(): if line.get_label() == '2023': line.set_linewidth(5) figure.get_figure().savefig('cumulative-minutes.png') return df
Weeks to CFP | 2019 | 2020 | 2021 | 2022 | 2023 |
---|---|---|---|---|---|
-12 | 0 | 0 | 0 | 0 | 70 |
-9 | 0 | 0 | 0 | 0 | 100 |
-8 | 0 | 0 | 50 | 25 | 100 |
-7 | 0 | 0 | 67 | 25 | 130 |
-6 | 0 | 0 | 74 | 25 | 130 |
-5 | 45 | 10 | 96 | 56 | 160 |
-4 | 66 | 25 | 115 | 56 | 220 |
-3 | 66 | 25 | 188 | 87 | 260 |
-2 | 192 | 55 | 198 | 104 | 390 |
-1 | 274 | 123 | 361 | 295 | 570 |
0 | 422 | 547 | 558 | 405 | 710 |
1 | 422 | 699 | 568 | 512 | 730 |
2 | 429 | 699 | 578 | 512 | 730 |
So… yeah… 730 minutes of talks for this year… I might've gotten a little carried away. But I like all the talks! And I want them to be captured in videos and maybe even transcribed by people who will take the time to change misrecognized words like Emax into Emacs! And I want people to be able to connect with other people who are interested in the sorts of stuff they're doing! So we're going to make it happen. The draft schedule's looking pretty full, but I think it'll work out, especially if the speakers send in their videos on time. Let's see how it all works out!
(…and look, I even got to learn how to do pivot tables and graphs with Python!)
]]>import logging from llama_index import GPTSimpleVectorIndex, MockLLMPredictor, MockEmbedding, QuestionAnswerPrompt logging.basicConfig(level=logging.DEBUG) logging.debug('This is a test.')
As it turns out, there was already a call to basicConfig in
github_repository_reader.py
which llama_index
loaded at some
point, and the Python documentation says: "As it’s intended as a
one-off simple configuration facility, only the first call will
actually do anything: subsequent calls are effectively no-ops."
So this is what I needed to do instead:
logging.getLogger().setLevel(logging.DEBUG)
yyyy-mm-dd-nn
to identify my
sketches. To avoid duplicates, I get these IDs from the web-based
journaling system I wrote. I've started putting the titles and tags
into those journal entries as well so that I can reuse them in
scripts. When I export a sketch to PNG and synchronize it, the file
appears in my ~/Dropbox/Supernote/EXPORT
directory on my laptop.
Then it goes through this process:
The following code does that processing.
#!/usr/bin/python3 # -*- mode: python -*- # (c) 2022-2023 Sacha Chua (sacha@sachachua.com) - MIT License # Permission is hereby granted, free of charge, to any person # obtaining a copy of this software and associated documentation files # (the "Software"), to deal in the Software without restriction, # including without limitation the rights to use, copy, modify, merge, # publish, distribute, sublicense, and/or sell copies of the Software, # and to permit persons to whom the Software is furnished to do so, # subject to the following conditions: # The above copyright notice and this permission notice shall be # included in all copies or substantial portions of the Software. # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, # EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF # MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND # NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS # BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN # ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN # CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE # SOFTWARE. import os import json import re import requests import time from dotenv import load_dotenv # Import the Google Cloud client libraries from google.cloud import vision from google.cloud.vision_v1 import AnnotateImageResponse import sys sys.path.append("/home/sacha/proj/supernote/") import recolor # noqa: E402 # muffles flake8 error about import load_dotenv() # Set the folder path where the png files are located folder_path = '/home/sacha/Dropbox/Supernote/EXPORT/' public_sketch_dir = '/home/sacha/sync/sketches/' private_sketch_dir = '/home/sacha/sync/private-sketches/' # Initialize the Google Cloud Vision client client = vision.ImageAnnotatorClient() refresh_counter = 0 def extract_text(client, file): json_file = file[:-3] + 'json' # TODO Preprocess to keep only black text with open(file, 'rb') as image_file: content = image_file.read() # Convert the png file to a Google Cloud Vision image object image = vision.Image(content=content) # Extract handwriting from the image using the Google Cloud Vision API response = client.document_text_detection(image=image) response_json = AnnotateImageResponse.to_json(response) json_response = json.loads(response_json) # Save the response to a json file with the same name as the png file with open(json_file, "w") as f: json.dump(json_response, f) def maybe_rename(file): # TODO Match on ID json_file = file[:-3] + 'json' with open(json_file, 'r') as f: data = json.load(f) # Extract the text from the json file text = data['fullTextAnnotation']['text'] # Check if the text contains a string matching the regex pattern pattern = r'(?<!ref:)[0-9]{4}-[0-9]{2}-[0-9]{2}-[0-9]{2}' match = re.search(pattern, text) if match: # Get the matched string matched_string = match.group(0) new_name = matched_string from_zid = get_journal_entry(matched_string).strip() if from_zid: new_name = matched_string + ' ' + from_zid tags = get_tags(new_name, text) if tags: new_name = new_name + ' ' + tags ref = get_references(text) if ref: new_name = new_name + ' ' + ref print('Renaming ' + file + ' to ' + new_name) # Rename the png and json files to the matched string new_filename = os.path.join(os.path.dirname(file), new_name + '.png') rename_set(file, new_filename) return new_filename def get_tags(filename, text): tags = re.findall(r'(^|\W)#[ \n\t]+', text) return ' '.join(filter(lambda x: x not in filename, tags)) def get_references(text): refs = re.findall(r'!ref:[0-9]{4}-[0-9]{2}-[0-9]{2}-[0-9]{2}', text) return ' '.join(refs) def get_journal_entry(zid): resp = requests.get('https://' + os.environ['JOURNAL_USER'] + ':' + os.environ['JOURNAL_PASS'] + '@journal.sachachua.com/api/entries/' + zid) j = resp.json() if j and not re.search('^I thought about', j['Note']): return j['Note'] def get_color_map(filename, text=None): if text: together = filename + ' ' + text else: together = filename if re.search('r#(parenting|purple|life)', together): return {'9d9d9d': '8754a1', 'c9c9c9': 'e4c1d9'} # parenting is purplish elif re.search(r'#(emacs|geek|tech|blue)', together): return {'9d9d9d': '2b64a9', 'c9c9c9': 'b3e3f1'} # geeky stuff in light/dark blue else: return {'9d9d9d': '884636', 'c9c9c9': 'f6f396'} # yellow highlighter, dark brown def rename_set(old_name, new_name): if old_name != new_name: old_json = old_name[:-3] + 'json' new_json = new_name[:-3] + 'json' os.rename(old_name, new_name) os.rename(old_json, new_json) def recolor_based_on_filename(filename): color_map = get_color_map(filename) recolored = recolor.map_colors(filename, color_map) # possibly rename based on the filename new_filename = re.sub(' #(purple|blue)', '', filename) rename_set(filename, new_filename) recolored.save(new_filename) def move_processed_sketch(file): global refresh_counter if '#private' in file: output_dir = private_sketch_dir elif '#' in file: output_dir = public_sketch_dir refresh_counter = 3 else: return file new_filename = os.path.join(output_dir, os.path.basename(file)) rename_set(file, new_filename) return new_filename def process_file(file): json_file = file[:-3] + 'json' # Check if a corresponding json file already exists if not os.path.exists(json_file): extract_text(client, file) if not re.search('[0-9]{4}-[0-9]{2}-[0-9]{2}-[0-9]{2} ', file): file = maybe_rename(file) recolor_based_on_filename(file) move_processed_sketch(file) def process_dir(folder_path): global processed_files # Iterate through all png files in the specified folder files = sorted(os.listdir(folder_path)) for file in files: if file.endswith('.png') and '_' in file: print("Processing ", file) process_file(os.path.join(folder_path, file)) def daemon(folder_path, wait): global refresh_counter while True: process_dir(folder_path) time.sleep(wait) if refresh_counter > 0: refresh_counter = refresh_counter - 1 if refresh_counter == 0: print("Reloading sketches") requests.get('https://' + os.environ['JOURNAL_USER'] + ':' + os.environ['JOURNAL_PASS'] + '@sketches.sachachua.com/reload?python=1') if __name__ == '__main__': # Create a set to store the names of processed files processed_files = set() if len(sys.argv) > 1: if os.path.isdir(sys.argv[1]): folder_path = sys.argv[1] daemon(folder_path, 300) else: for f in sys.argv[1:]: process_file(f) else: daemon(folder_path, 300)
It uses this script I wrote to recolor my sketches with Python.
I'm contemplating writing some annotation tools to make it easier to turn the detected text into useful text for searching or writing about because the sketches throw off the recognition (misrecognized text, low confidence) and the columns mess up the line wrapping. Low priority, though.
My handwriting (at least for numbers) is probably simple enough that I might be able to train Tesseract OCR to process that someday. And who knows, maybe some organization will release a pre-trained model for offline handwriting recognition that'll be as useful as OpenAI Whisper is for audio files. That would be neat!
]]>: Removed my fork since upstream now has the :eval function.
The Q&A session for Things I'd like to see in Emacs (Richard Stallman) from EmacsConf 2022 was done over Mumble. Amin pasted the questions into the Mumble chat buffer and I copied them into a larger buffer as the speaker answered them, but I didn't do it consistently. I figured it might be worth making another video with easier-to-read visuals. At first, I thought about using LaTeX to create Beamer slides with the question text, which I could then turn into a video using ffmpeg. Then I decided to figure out how to animate the text in Emacs, because why not? I figured a straightforward typing animation would probably be less distracting than animate-string
, and emacs-director seems to handle that nicely. I forked it to add a few things I wanted, like variables to make the typing speed slower (so that it could more reliably type things on my old laptop, since sometimes the timers seemed to have hiccups) and an . (2023-01-14: Upstream has the :eval feature now.)
:eval
step for running things without needing to log them
To make it easy to synchronize the resulting animation with the chapter markers I derived from the transcript of the audio file, I decided to beep between scenes. First step: make a beep file.
ffmpeg -y -f lavfi -i 'sine=frequency=1000:duration=0.1' beep.wav
Next, I animated the text, with a beep between scenes. I used
subed-parse-file
to read the question text directly from the chapter
markers, and I used simplescreenrecorder to set up the recording
settings (including audio).
(defun my-beep () (interactive) (save-window-excursion (shell-command "aplay ~/recordings/beep.wav &" nil nil))) (require 'director) (defvar emacsconf-recording-process nil) (shell-command "xdotool getwindowfocus windowsize 1282 720") (progn (switch-to-buffer (get-buffer-create "*Questions*")) (erase-buffer) (org-mode) (face-remap-add-relative 'default :height 300) (setq-local mode-line-format " Q&A for EmacsConf 2022: What I'd like to see in Emacs (Richard M. Stallman) - emacsconf.org/2022/talks/rms") (sit-for 3) (delete-other-windows) (hl-line-mode -1) (when (process-live-p emacsconf-recording-process) (kill-process emacsconf-recording-process)) (setq emacsconf-recording-process (start-process "ssr" (get-buffer-create "*ssr*") "simplescreenrecorder" "--start-recording" "--start-hidden")) (sit-for 3) (director-run :version 1 :log-target '(file . "/tmp/director.log") :before-start (lambda () (switch-to-buffer (get-buffer-create "*Questions*")) (delete-other-windows)) :steps (let ((subtitles (subed-parse-file "~/proj/emacsconf/rms/emacsconf-2022-rms--what-id-like-to-see-in-emacs--answers--chapters.vtt"))) (apply #'append (list (list :eval '(my-beep)) (list :type "* Q&A for Richard Stallman's EmacsConf 2022 talk: What I'd like to see in Emacs\nhttps://emacsconf.org/2022/talks/rms\n\n")) (mapcar (lambda (sub) (list (list :log (elt sub 3)) (list :eval '(progn (org-end-of-subtree) (unless (bolp) (insert "\n")))) (list :type (concat "** " (elt sub 3) "\n\n")) (list :eval '(org-back-to-heading)) (list :wait 5) (list :eval '(my-beep)))) subtitles))) :typing-style 'human :delay-between-steps 0 :after-end (lambda () (process-send-string emacsconf-recording-process "record-save\nwindow-show\nquit\n")) :on-failure (lambda () (process-send-string emacsconf-recording-process "record-save\nwindow-show\nquit\n")) :on-error (lambda () (process-send-string emacsconf-recording-process "record-save\nwindow-show\nquit\n"))))
I used the following code to copy the latest recording to animation.webm
and extract the audio to animation.wav
. my-latest-file
and my-recordings-dir
are in my Emacs config.
(let ((name "animation.webm")) (copy-file (my-latest-file my-recordings-dir) name t) (shell-command (format "ffmpeg -y -i %s -ar 8000 -ac 1 %s.wav" (shell-quote-argument name) (shell-quote-argument (file-name-sans-extension name)))))
Then I needed to get the timestamps of the beeps in the recording. I subtracted a little bit (0.82
seconds) based on comparing the waveform with the results.
filename = "animation.wav" from scipy.io import wavfile from scipy import signal import numpy as np import re rate, source = wavfile.read(filename) peaks = signal.find_peaks(source, height=1000, distance=1000) base_times = (peaks[0] / rate) - 0.82 print(base_times)
I noticed that the first question didn't seem to get beeped properly, so I tweaked the times. Then I wrote some code to generate a very long ffmpeg command that used trim and tpad to select the segments and extend them to the right durations. There was some drift when I did it without the audio track, but the timestamps seemed to work right when I included the Q&A audio track as well.
import webvtt import subprocess chapters_filename = "emacsconf-2022-rms--what-id-like-to-see-in-emacs--answers--chapters.vtt" answers_filename = "answers.wav" animation_filename = "animation.webm" def get_length(filename): result = subprocess.run(["ffprobe", "-v", "error", "-show_entries", "format=duration", "-of", "default=noprint_wrappers=1:nokey=1", filename], stdout=subprocess.PIPE, stderr=subprocess.STDOUT) return float(result.stdout) def get_frames(filename): result = subprocess.run(["ffprobe", "-v", "error", "-select_streams", "v:0", "-count_packets", "-show_entries", "stream=nb_read_packets", "-of", "csv=p=0", filename], stdout=subprocess.PIPE, stderr=subprocess.STDOUT) return float(result.stdout) answers_length = get_length(answers_filename) # override base_times times = np.asarray([ 1.515875, 13.50, 52.32125 , 81.368625, 116.66625 , 146.023125, 161.904875, 182.820875, 209.92125 , 226.51525 , 247.93875 , 260.971 , 270.87375 , 278.23325 , 303.166875, 327.44925 , 351.616375, 372.39525 , 394.246625, 409.36325 , 420.527875, 431.854 , 440.608625, 473.86825 , 488.539 , 518.751875, 544.1515 , 555.006 , 576.89225 , 598.157375, 627.795125, 647.187125, 661.10875 , 695.87175 , 709.750125, 717.359875]) fps = 30.0 times = np.append(times, get_length(animation_filename)) anim_spans = list(zip(times[:-1], times[1:])) chapters = webvtt.read(chapters_filename) if chapters[0].start_in_seconds == 0: vtt_times = [[c.start_in_seconds, c.text] for c in chapters] else: vtt_times = [[0, "Introduction"]] + [[c.start_in_seconds, c.text] for c in chapters] vtt_times = vtt_times + [[answers_length, "End"]] # Add ending timestamps vtt_times = [[x[0][0], x[1][0], x[0][1]] for x in zip(vtt_times[:-1], vtt_times[1:])] test_rate = 1.0 i = 0 concat_list = "" groups = list(zip(anim_spans, vtt_times)) import ffmpeg animation = ffmpeg.input('animation.webm').video audio = ffmpeg.input('rms.opus') for_overlay = ffmpeg.input('color=color=black:size=1280x720:d=%f' % answers_length, f='lavfi') params = {"b:v": "1k", "vcodec": "libvpx", "r": "30", "crf": "63"} test_limit = 1 params = {"vcodec": "libvpx", "r": "30", "copyts": None, "b:v": "1M", "crf": 24} test_limit = 0 anim_rate = 1 import math cursor = 0 if test_limit > 0: groups = groups[0:test_limit] clips = [] # cursor is the current time for anim, vtt in groups: padding = vtt[1] - cursor - (anim[1] - anim[0]) / anim_rate if (padding < 0): print("Squeezing", math.floor((anim[1] - anim[0]) / (anim_rate * 1.0)), 'into', vtt[1] - cursor, padding) clips.append(animation.trim(start=anim[0], end=anim[1]).setpts('PTS-STARTPTS')) elif padding == 0: clips.append(animation.trim(start=anim[0], end=anim[1]).setpts('PTS-STARTPTS')) else: print("%f to %f: Padding %f into %f - pad: %f" % (cursor, vtt[1], (anim[1] - anim[0]) / (anim_rate * 1.0), vtt[1] - cursor, padding)) cursor = cursor + padding + (anim[1] - anim[0]) / anim_rate clips.append(animation.trim(start=anim[0], end=anim[1]).setpts('PTS-STARTPTS').filter('tpad', stop_mode="clone", stop_duration=padding)) for_overlay = for_overlay.overlay(animation.trim(start=anim[0], end=anim[1]).setpts('PTS-STARTPTS+%f' % vtt[0])) clips.append(audio.filter('atrim', start=vtt[0], end=vtt[1]).filter('asetpts', 'PTS-STARTPTS')) args = ffmpeg.concat(*clips, v=1, a=1).output('output.webm', **params).overwrite_output().compile() print(' '.join(f'"{item}"' for item in args))
Anyway, it's here for future reference. =)
]]>YYYY-MM-DD HH:MM:SS
and someone's computer decided to turn it into
MM/DD/YY HH:MM
. To avoid this conversion and import the columns as
strings, you can change the file extension to .txt
instead of .csv
and then change each column type that you care about, which can be a
lot of clicking. I had to change things back with a regular expression
along the lines of:
import re s = "12/9/21 11:23" match = re.match('([0-9]+)/([0-9]+)/([0-9]+)( [0-9]+:[0-9]+)', s) date = '20%s-%s-%s%s:00' % (match.group(3).zfill(2), match.group(1).zfill(2), match.group(2).zfill(2), match.group(4)) print(date)
The pandas
library for Python also likes to do this kind of data
type conversion for data types and for NaN values. In this particular
situation, I wanted it to leave columns alone and leave the nan
string in my input alone. Otherwise, to_csv
would replace nan
with
the blank string, which could mess up a different script that used
this data as input. This is the code to do it:
import pandas as pd df = pd.read_csv('filename.csv', encoding='utf-8', dtype=str, na_filter=False)
I'm probably going to run into this again sometime, so I wanted to make sure I put my notes somewhere I can find them later.
]]>