add script for adding keynoters and keynoters images

docs/keynoter-scheduling.md

@@ -0,0 +1,144 @@
+# Keynoter Social Media Scheduling
+
+This document describes the process for scheduling keynote speaker announcement
+posts to Buffer. In short, you should:
+
+- prepare the images in canva and deploy them to have a live url to the image
+- prepare copy for each platform and put it into a json file
+- run the script to schedule posts for the keynoters, one at a time, to all the
+  supported social media platform
+
+The script posts to all connected channels in one run per keynoter: `instagram`,
+`linkedin`, `fosstodon`, `bsky`, `x`, `tiktok`.
+
+---
+
+## Prerequisites
+
+- Python environment with `requests` and `python-dotenv` installed
+- A `.env.local` file in the **repo root** with your Buffer API key:
+
+```
+BUFFER_API_KEY=your_buffer_api_key_here
+```
+
+### Getting the Buffer API key
+
+1. Log in to [buffer.com](https://buffer.com) with the EuroPython account
+2. Go to **Account Settings → Apps & Integrations**
+3. Copy the personal access token under **Access Token**
+4. Paste it as `BUFFER_API_KEY` in `.env.local`
+
+---
+
+## Step 1 — Prepare and deploy the images
+
+Buffer requires a **publicly accessible image URL** — local file paths, Google
+Drive links, and Canva share links do not work. The image must be hosted on the
+live site before the script can use it.
+
+### Prepare the images
+
+- Export images as PNG from your design tool (Canva)
+- Name them `firstname-lastname.png` (lowercase, hyphenated)
+- Place them in `website/public/media/keynoters/`
+
+### Deploy
+
+1. Commit the images and open a PR
+2. Wait for the PR to be merged and deployed to production
+3. Verify each image is accessible, e.g.:
+   `https://ep2026.europython.eu/media/keynoters/leah-wasser.png`
+
+> ⚠️ Do not run the scheduling script until the images are live. Buffer fetches
+> the URL at scheduling time and will fail with a "Not Found" error if the file
+> hasn't been deployed yet.
+
+## Step 2 — Prepare the post copy (`keynoters.json`)
+
+All post text lives in `scripts/keynoters.json`
+
+```json
+{
+  "Speaker Name": {
+    "image": "https://ep<year>.europython.eu/media/keynoters/firstname-lastname.png",
+    "instagram": "Post text for Instagram...",
+    "linkedin": "Post text for LinkedIn...",
+    "fosstodon": "Post text for Fosstodon/Mastodon...",
+    "bsky": "Post text for Bluesky...",
+    "x": "Post text for X/Twitter...",
+    "tiktok": "Post text for TikTok..."
+  }
+}
+```
+
+A few things to keep in mind when writing copy:
+
+- Social handles (e.g. `@leahawasser.bsky.social`) go **inline in the post
+  text**, not as separate fields. Each platform's post should use the handle
+  format native to that platform.
+- X and Bluesky have character limits — keep those posts short.
+- Instagram and TikTok don't render clickable URLs, so use the short form
+  (`europython.eu/tickets/`) rather than the full URL.
+- LinkedIn and Fosstodon support full clickable URLs.
+
+Add one entry per keynoter. The key must match exactly what you'll set in the
+script in Step 3.
+
+---
+
+## Step 3 — Schedule posts via Buffer
+
+The scheduling script is `scripts/buffer-keynoters.py`. Run it once per
+keynoter.
+
+### Configure the script
+
+Open `scripts/buffer-keynoters.py` and edit the two lines at the top:
+
+```python
+KEYNOTER     = "Leah Wasser"           # must match the key in keynoters.json exactly
+SCHEDULED_AT = datetime(2026, 6, 16, 10, 0,
+                        tzinfo=ZoneInfo("Europe/London"))
+```
+
+Set `SCHEDULED_AT` to the date and time you want the post to go live. The
+timezone is `Europe/London` — adjust the year, month, day, hour, and minute as
+needed.
+
+### Preview before posting
+
+```bash
+python scripts/buffer-keynoters.py --dry-run
+```
+
+This prints the first 200 characters of the post for each platform without
+sending anything to Buffer. Use it to sanity-check the copy and confirm the
+right keynoter is selected.
+
+### Run
+
+```bash
+python scripts/buffer-keynoters.py
+```
+
+The script will:
+
+1. Connect to Buffer and fetch your channel IDs
+2. Fetch the image from the live URL in `keynoters.json`
+3. Schedule the post to every platform that has text defined
+4. Skip any platform not connected in your Buffer account
+5. Print a confirmation line with the Buffer post ID for each channel
+
+Repeat for each keynoter, updating `KEYNOTER` and `SCHEDULED_AT` each time.
+
+---
+
+## Troubleshooting
+
+| Error                               | Cause                        | Fix                                              |
+| ----------------------------------- | ---------------------------- | ------------------------------------------------ |
+| `BUFFER_API_KEY not set`            | Missing `.env.local`         | Create `website/.env.local` with the key         |
+| `Image upload failed (404)`         | Image not deployed yet       | Merge the images PR and wait for deployment      |
+| `not connected in Buffer — skipped` | Channel not linked in Buffer | Log in to Buffer and connect the missing channel |
+| `not found in keynoters.json`       | Typo in `KEYNOTER`           | Check the exact key spelling in `keynoters.json` |

docs/social-media-scheduling.md

@@ -109,24 +109,24 @@ speaker → speaker → sponsor → speaker → partner → (repeat)
 Regenerate it by hitting the API endpoint while the dev server is running:
 
 ```bash
-curl -s http://localhost:4321/api/media/combined_socials_queue > src/pages/api/media/combined_socials_queue.json
+curl -s http://localhost:4321/api/media/combined_socials_queue > scripts/combined_socials_queue.json
 ```
 
-This overwrites `src/pages/api/media/combined_socials_queue.json` with all items
-sorted and interleaved.
+This overwrites `scripts/combined_socials_queue.json` with all items sorted and
+interleaved.
 
 > **Note:** There are two queue files:
 >
-> - `combined_socials_queue.json` — the freshly generated queue, used as the
->   source of truth and read by `buffer-scheduling.py`
-> - `combined_socials_queue_2026.json` — the manually curated queue that
+> - `scripts/combined_socials_queue.json` — the freshly generated queue, used as
+>   the source of truth and read by `buffer-scheduling.py`
+> - `scripts/combined_socials_queue_2026.json` — the manually curated queue that
 >   preserves already-scheduled posts at the top; new entries are appended after
 >   the last scheduled position
 
 When new speakers or sponsors are added, regenerate
-`combined_socials_queue.json` and then merge the new entries into
-`combined_socials_queue_2026.json` manually — keeping already-posted entries
-intact and appending only new/unposted ones.
+`scripts/combined_socials_queue.json` and then merge the new entries into
+`scripts/combined_socials_queue_2026.json` manually — keeping already-posted
+entries intact and appending only new/unposted ones.
 
 ### Session types and labels
 
@@ -150,7 +150,8 @@ Sponsors are split into two buckets:
   Financial Aid
 
 If a new sponsor tier is added, update `commercialTiers` in
-`src/pages/api/media/combined_socials_queue.ts`.
+`src/pages/api/media/combined_socials_queue.ts` (this file stays in the API
+layer).
 
 ### Manual queue adjustments
 
@@ -160,7 +161,7 @@ reorder entries. For example, to swap two sponsors:
 ```python
 python3 -c "
 import json
-path = 'src/pages/api/media/combined_socials_queue.json'
+path = 'scripts/combined_socials_queue.json'
 with open(path) as f:
     q = json.load(f)
 q[2], q[17] = q[17], q[2]  # swap positions 3 and 18 (0-indexed)
@@ -169,9 +170,9 @@ with open(path, 'w') as f:
 "
 ```
 
-`src/pages/api/media/combined_socials_queue.json` is committed to the repo. Any
-manual reordering should be committed so the intentional order is preserved and
-not lost when the queue is regenerated.
+`scripts/combined_socials_queue.json` is committed to the repo. Any manual
+reordering should be committed so the intentional order is preserved and not
+lost when the queue is regenerated.
 
 ---
 
@@ -233,7 +234,7 @@ Buffer operates as a FIFO queue against pre-configured time slots:
 > remove them one by one. Before each run, open Buffer and verify the number and
 > cadence of slots for each channel matches your intended rollout pace.
 
-The scheduling script is at `src/pages/api/media/buffer-scheduling.py`.
+The scheduling script is at `scripts/buffer-scheduling.py`.
 
 ### Configuration
 
@@ -251,13 +252,13 @@ the results, increment the range for subsequent runs.
 ### Running the script
 
 ```bash
-uv run python src/pages/api/media/buffer-scheduling.py
+uv run python scripts/buffer-scheduling.py
 ```
 
 Or with your venv activated:
 
 ```bash
-python src/pages/api/media/buffer-scheduling.py
+python scripts/buffer-scheduling.py
 ```
 
 ### What it does
@@ -292,9 +293,9 @@ python src/pages/api/media/buffer-scheduling.py
 After each successful run, note the last `QUEUE_END` value. The next run should
 set `QUEUE_START = previous QUEUE_END + 1`.
 
-The script reads from `combined_socials_queue.json` by default. When switching
-to the curated `combined_socials_queue_2026.json`, update `queue_path` in the
-script accordingly.
+The script reads from `scripts/combined_socials_queue.json` by default. When
+switching to the curated `scripts/combined_socials_queue_2026.json`, update
+`queue_path` in the script accordingly.
 
 Already scheduled as of June 2026:
 

scripts/requirements.txt

@@ -0,0 +1,2 @@
+requests
+python-dotenv