From 4821062de1e2f18240227db0b9b359ee5fcae4cd Mon Sep 17 00:00:00 2001
From: Alejandro de la Vega <aleph4@gmail.com>
Date: Mon, 25 Aug 2025 12:45:35 -0500
Subject: [PATCH 1/3] Fix: PMCID extraction

---
 src/pubget/_utils.py | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/src/pubget/_utils.py b/src/pubget/_utils.py
index a1c88a0..b94fe97 100644
--- a/src/pubget/_utils.py
+++ b/src/pubget/_utils.py
@@ -136,12 +136,14 @@ def get_pmcid(article: Union[etree.ElementTree, etree.Element]) -> int:
     pmcid = article.find(
         "front/article-meta/article-id[@pub-id-type='pmcid']"
     )
-    if pmc is None and pmcid is None:
-        raise ValueError("No PMC ID found in the article XML.")
-    if pmc:
+
+    val = None
+    if pmc is not None:
         val = pmc.text
-    else:
-        val = pmcid.text.replace("PMC", "")
+    elif pmcid is not None:
+        val = pmcid.text
+        if val.startswith("PMC"):
+            val = val[3:]
 
     return int(val)
 

From 755d92eac5e779ec961a9354e8a636d7252fb5e6 Mon Sep 17 00:00:00 2001
From: Alejandro de la Vega <aleph4@gmail.com>
Date: Mon, 25 Aug 2025 12:51:50 -0500
Subject: [PATCH 2/3] Raise error if val is still None

---
 src/pubget/_utils.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/src/pubget/_utils.py b/src/pubget/_utils.py
index b94fe97..a50140d 100644
--- a/src/pubget/_utils.py
+++ b/src/pubget/_utils.py
@@ -145,6 +145,9 @@ def get_pmcid(article: Union[etree.ElementTree, etree.Element]) -> int:
         if val.startswith("PMC"):
             val = val[3:]
 
+    if val is None or not val.isdigit():
+        raise ValueError("No valid PMCID found in article XML.")
+
     return int(val)
 
 

From 1df08d227455ed9b66912a10e99d9e9dd09936a8 Mon Sep 17 00:00:00 2001
From: Alejandro de la Vega <aleph4@gmail.com>
Date: Fri, 29 Aug 2025 14:49:17 -0500
Subject: [PATCH 3/3] Fix PMCID extraction from text

---
 src/pubget/_metadata.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/pubget/_metadata.py b/src/pubget/_metadata.py
index 8728f2e..abbdcab 100644
--- a/src/pubget/_metadata.py
+++ b/src/pubget/_metadata.py
@@ -90,7 +90,7 @@ def _add_license(article: etree.Element, metadata: Dict[str, Any]) -> None:
 
 def _add_id(article_id: etree.Element, metadata: Dict[str, Any]) -> None:
     id_type = article_id.get("pub-id-type")
-    if id_type not in ["pmc", "pmid", "doi"]:
+    if id_type not in ["pmc", "pmid", "doi", "pmcid"]:
         return
     if id_type == "pmc":
         id_type = "pmcid"