From 0427602f4cf26b196f418f537cf681d273ea497d Mon Sep 17 00:00:00 2001
From: Daniel Obraczka <daniel@obraczka.de>
Date: Thu, 11 Apr 2024 11:17:03 +0200
Subject: [PATCH] Fallback use for existing cluster helper file if no parquet
 is present (#41)

* Use fallback for cache loading if cluster helper file exists

* Use constant for column names
---
 sylloge/base.py | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/sylloge/base.py b/sylloge/base.py
index 3c2c441..b6c2c90 100644
--- a/sylloge/base.py
+++ b/sylloge/base.py
@@ -310,6 +310,16 @@ def _read_ch_or_df_links(
             raise ValueError(
                 "Need to supply read_parquet_fn if not using ClusterHelper"
             )
+        parquet_path = f"{path}_parquet"
+        if not os.path.exists(parquet_path) and os.path.exists(path):
+            logger.info(
+                f"Did not find {parquet_path}, but ClusterHelper file. Creating parquet file from ClusterHelper (no intra-dataset links are used!)"
+            )
+            assert ds_prefixes is not None
+            ch = PrefixedClusterHelper.from_file(path, ds_prefixes=ds_prefixes)  # type: ignore[return-value]
+            pd.DataFrame(list(ch.all_pairs_no_intra()), columns=EA_SIDES).to_parquet(
+                parquet_path
+            )
         return read_parquet_fn(f"{path}_parquet", **kwargs)  # type: ignore[return-value]
 
     @classmethod