From 6a1f3e0d5612a6f7d0d2dccfe01135404aca3194 Mon Sep 17 00:00:00 2001
From: "Jurriaan H. Spaaks" <j.spaaks@esciencecenter.nl>
Date: Wed, 2 Feb 2022 12:33:10 +0100
Subject: [PATCH 1/2] added cherrypicked repositories

---
 cherrypicked.json | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)
 create mode 100644 cherrypicked.json

diff --git a/cherrypicked.json b/cherrypicked.json
new file mode 100644
index 0000000..8eeb3f9
--- /dev/null
+++ b/cherrypicked.json
@@ -0,0 +1,29 @@
+[
+    "https://github.com/allenai/allennlp",
+    "https://github.com/automl/Auto-PyTorch",
+    "https://github.com/automl/auto-sklearn",
+    "https://github.com/duckdb/duckdb",
+    "https://github.com/effector/effector",
+    "https://github.com/explosion/spaCy",
+    "https://github.com/github/linguist",
+    "https://github.com/huggingface/transformers",
+    "https://github.com/JuliaLang/julia",
+    "https://github.com/lutzroeder/netron",
+    "https://github.com/ManimCommunity/manim",
+    "https://github.com/microsoft/FLAML",
+    "https://github.com/microsoft/nni",
+    "https://github.com/microsoft/onnxruntime",
+    "https://github.com/mxgmn/WaveFunctionCollapse",
+    "https://github.com/nextflow-io/nextflow",
+    "https://github.com/nlohmann/json",
+    "https://github.com/nltk/nltk",
+    "https://github.com/open-mmlab/mmdetection",
+    "https://github.com/pydata/xarray",
+    "https://github.com/PyTorchLightning/pytorch-lightning",
+    "https://github.com/RaRe-Technologies/gensim",
+    "https://github.com/sagemath/sage",
+    "https://github.com/spack/spack",
+    "https://github.com/tensorflow/federated",
+    "https://github.com/tensorflow/tensorflow",
+    "https://github.com/Zettlr/Zettlr"
+]

From c67f894777c48be6dab298d14c7b00958d9858d3 Mon Sep 17 00:00:00 2001
From: "Jurriaan H. Spaaks" <j.spaaks@esciencecenter.nl>
Date: Wed, 2 Feb 2022 12:33:46 +0100
Subject: [PATCH 2/2] now using cherrypicked list of repos as starting point
 for filtering

---
 query.mjs | 15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

diff --git a/query.mjs b/query.mjs
index 8f76e45..a223fbd 100644
--- a/query.mjs
+++ b/query.mjs
@@ -149,8 +149,6 @@ const filterAsync = async (arr, asyncCallback) => {
 }
 
 
-const urls_rsd = loadFromJsonfile('./urls.json');
-
 const nworkflows_minimum = 1;
 const npull_requests_minimum = 5;
 const inactivity_threshold = 12 * 30 * 24 * 60 * 60 * 1000 // X months in milliseconds -> X months * 30 days/month * 24 hours/day * 60 min/hour * 60 sec/min * 1000
@@ -159,21 +157,12 @@ const ncontributions_minimum = 5;
 
 const octokit = new Octokit({auth: process.env.GITHUB_TOKEN});
 
-let urls = urls_rsd;
+let urls = loadFromJsonfile('./cherrypicked.json');
 urls = await filterAsync(urls, includeHasCitationcff);
 urls = await filterAsync(urls, includeUsesPullRequests);
 urls = await filterAsync(urls, hasMultipleChangesToCitationcff);
 urls = await filterAsync(urls, includeUsesWorkflows);
 urls = await filterAsync(urls, hasRecentCommits);
-urls = await filterAsync(urls, includeHasValidcff);
 urls = await filterAsync(urls, hasSufficientContributors);
+urls = await filterAsync(urls, includeHasValidcff);
 urls.forEach(url => console.log(url))
-
-
-//const q = 'cffconvert-github-action in:file path:.github/workflows';
-// const q = 'CITATION.cff in:path path:/';
-// const { data } = await octokit.request('GET /search/code', { q });
-
-//data.items.forEach((item) => {
-//  console.log(item.repository.html_url);
-//})