Consolidating Daily Content Emails: From Platform‑Specific Sends to a Unified Apple‑Style Summary

<h1> Consolidating Daily Content Emails: From Platform‑Specific Sends to a Unified Apple‑Style Summary </h1> TL;DR: We replaced multiple platform‑specific email jobs with a single, metadata‑driven consolidator that builds one Apple‑style summary. The change cut duplication, eliminated missing‑field errors, and made the daily workflow easier to maintain. <h2> The Problem </h2> In our previous setup, each platform (Bluesky, Dev.to, Substack, Medium, etc.) had its own GitHub Action that triggered <code>src/notifier.py</code> to send a daily email. The email body was built directly from the platform’s content files. This architecture caused two major issues: <ol> <li> Duplication of effort – Every platform required its own email template, leading to 6 separate code paths that almost always stayed in sync.</li> <li> Fragile metadata – The email generation relied on the presence of specific keys in <code>metadata.json</code>. When a post missed the <code>bluesky_uris</code> field, the notifier raised a <code>KeyError</code>, aborting the entire daily send: </li> </ol> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code> Traceback (most recent call last): File "src/notifier.py", line 112, in send_summary bluesky_cell = f'<a href="{metadata["bluesky_uris"]["bluesky"]}">…' KeyError: 'bluesky_uris' </code></pre> </div> The result was a flaky daily workflow that required manual fixes and manual updates to the email templates whenever the content structure changed. <h2> What I Tried First </h2> Initially I attempted a refactor of each individual workflow: <ul> <li>Updated each GitHub Action to use a shared <code>notifier</code> library.</li> <li>Centralized the email template but kept separate <code>send_*</code> functions for each platform.</li> </ul> This approach still required maintaining six distinct functions and duplicated logic for reading metadata, constructing links, and handling missing fields. The code still broke when a single post was missing a field for one platform. <h2> The Implementation </h2> <h3> 1. Centralized Metadata Reader </h3> I created <code>src/daily_consolidator.py</code> to read all content files for a given day and aggregate the necessary data into a single dictionary: <div class="highlight js-code-highlight"> <pre class="highlight python"><code># src/daily_consolidator.py import json import os from pathlib import Path def load_metadata(date: str) -> dict: """Load all metadata for a given date and merge them.""" base = Path(f"content/{date}") meta_files = list(base.glob("*/metadata.json")) merged = {} for <span class=