Add WELOPEN support in ecl2df COMPDAT parsing (#186)

* Add WELOPEN support in ecl2df COMPDAT parsing

* Closes #185

* Drop f-string for py2 compatibility

* Change into a function

* Add test for applywelopen

* Improved readability

* Fix nan to None

* Add test with defined IJK WELOPEN

* Type welLopen

* Warning if lumped connections

* Add test for invalid WELOPEN well

* Move test, remove unneeded columns

* Removed varitation in indentation test sim decks

* Add comment for column name translation

* Updated docstring for applywelopen

* Add WELOPEN to bash script

ecl2df/common.py

@@ -54,6 +54,7 @@
     "WCONINJE",
     "WCONINJH",
     "WCONPROD",
+    "WELOPEN",
     "WELSEGS",
     "WELSPECS",
 ]:

ecl2df/compdat.py

@@ -67,6 +67,7 @@ def deck2dfs(deck, start_date=None, unroll=True):
     """
     compdatrecords = []  # List of dicts of every line in input file
     compsegsrecords = []
+    welopenrecords = []
     welsegsrecords = []
     date = start_date  # DATE column will always be there, but can contain NaN/None
     for kword in deck:
@@ -105,6 +106,18 @@ def deck2dfs(deck, start_date=None, unroll=True):
                 rec_data["WELL"] = wellname
                 rec_data["DATE"] = date
                 compsegsrecords.append(rec_data)
+        elif kword.name == "WELOPEN":
+            for rec in kword:
+                rec_data = parse_opmio_deckrecord(rec, "WELOPEN")
+                rec_data["DATE"] = date
+                if rec_data["STATUS"] not in ["OPEN", "SHUT", "STOP", "AUTO"]:
+                    rec_data["STATUS"] = "SHUT"
+                    logger.warning(
+                        "WELOPEN status %s is not a valid "
+                        "COMPDAT state. Using 'SHUT' instead." % rec_data["STATUS"]
+                    )
+                welopenrecords.append(rec_data)
+
         elif kword.name == "WELSEGS":
             # First record contains meta-information for well
             # (opm deck returns default values for unspecified items.)
@@ -126,18 +139,19 @@ def deck2dfs(deck, start_date=None, unroll=True):
                 if "INFO_TYPE" in rec_data and rec_data["INFO_TYPE"] == "ABS":
                     rec_data["SEGMENT_MD"] = rec_data["SEGMENT_LENGTH"]
                 welsegsrecords.append(rec_data)
-        elif kword.name == "TSTEP":
-            logger.warning("Possible premature stop at first TSTEP")
-            break
 
     compdat_df = pd.DataFrame(compdatrecords)
+    welopen_df = pd.DataFrame(welopenrecords)
 
     if unroll and not compdat_df.empty:
         compdat_df = unrolldf(compdat_df, "K1", "K2")
 
-    compsegs_df = pd.DataFrame(compsegsrecords)
+    if not welopen_df.empty:
+        compdat_df = applywelopen(compdat_df, welopen_df)
 
+    compsegs_df = pd.DataFrame(compsegsrecords)
     welsegs_df = pd.DataFrame(welsegsrecords)
+
     if unroll and not welsegs_df.empty:
         welsegs_df = unrolldf(welsegs_df, "SEGMENT1", "SEGMENT2")
 
@@ -241,6 +255,99 @@ def unrolldf(dframe, start_column="K1", end_column="K2"):
     return unrolled
 
 
+def applywelopen(compdat_df, welopen_df):
+    """Apply WELOPEN actions to the COMPDAT dataframe.
+
+    Each record in the WELOPEN keyword acts as an operator on existing connections
+    in existing wells.
+
+    Example: COMPDAT and WELOPEN keyword::
+
+      COMPDAT
+       'OP1' 33 44 10 11 'OPEN' /
+       'OP2' 66 44 10 11 'OPEN' /
+      /
+      WELOPEN
+       'OP1' SHUT /
+       'OP2' SHUT 66 44 10 /
+      /
+
+    This deck would define two wells where OP1 and OP2 have two connected grid cells
+    each. Although the COMPDAT defines all connections to be open, WELOPEN overwrites
+    this: all connections in OP1 will be SHUT and in OP2 the upper connection will
+    be SHUT.
+
+    WELOPEN can also be used at different dates and changes therefore the state of
+    connections without explicit use of the COMPDAT keyword. This function translates
+    WELOPEN actions into explicit additional COMPDAT definitions in the exported df.
+
+    Args:
+        compdat_df (pd.DataFrame): Dataframe with unrolled COMPDAT data
+        welopen_df (pd.DataFrame): Dataframe with WELOPEN actions
+
+    Returns:
+        pd.Dataframe, compdat_df now including WELOPEN actions
+
+    """
+    welopen_df = welopen_df.astype(object).where(pd.notnull(welopen_df), None)
+    for _, row in welopen_df.iterrows():
+        if row["I"] and row["J"] and row["K"]:
+            previous_state = compdat_df[
+                (compdat_df["WELL"] == row["WELL"])
+                & (compdat_df["DATE"] <= row["DATE"])
+                & (compdat_df["I"] == row["I"])
+                & (compdat_df["J"] == row["J"])
+                & (compdat_df["K1"] == row["K"])
+                & (compdat_df["K2"] == row["K"])
+            ].drop_duplicates(subset=["I", "J", "K1", "K2"], keep="last")
+        elif row["C1"] or row["C2"]:
+            logger.warning(
+                "Lumped connections are not supported in a WELOPEN keyword. "
+                "Skipping WELOPEN actions for lumped connections '%s' and/or '%s'"
+                % (row["C1"], row["C2"])
+            )
+            continue
+        elif not (row["I"] and row["J"] and row["K"]):
+            previous_state = compdat_df[
+                (compdat_df["WELL"] == row["WELL"])
+                & (compdat_df["DATE"] <= row["DATE"])
+            ].drop_duplicates(subset=["I", "J", "K1", "K2"], keep="last")
+        else:
+            raise ValueError(
+                "A WELOPEN keyword contains data that could not be parsed. "
+                "(I=%s,J=%s,K=%s)" % (row["I"], row["J"], row["K"])
+            )
+
+        if previous_state.empty:
+            raise ValueError(
+                "A WELOPEN keyword is not acting on any existing connection. "
+                "(I=%s,J=%s,K=%s)" % (row["I"], row["J"], row["K"])
+            )
+
+        new_state = previous_state
+
+        # The COMPDAT DataFrame uses COMPDAT_RENAMER and therefore uses "OP/SH" as a
+        # column name for the state of a well. WELOPEN uses "STATUS" for the state
+        # column name and therefore a translation step needs to be done. The
+        # underlying problem is that the opm-common definitions for the state of a
+        # well in COMPDAT and WELOPEN are not identical. These translation steps can
+        # be dropped when unity in the opm-common keyword definitions is reached.
+        new_state["OP/SH"] = row["STATUS"]
+
+        new_state["DATE"] = row["DATE"]
+
+        compdat_df = compdat_df.append(new_state)
+
+    if not compdat_df.empty:
+        compdat_df = (
+            compdat_df.sort_values(by=["DATE", "WELL"])
+            .drop_duplicates(subset=["I", "J", "K1", "K2", "DATE"], keep="last")
+            .reset_index(drop=True)
+        )
+
+    return compdat_df
+
+
 def fill_parser(parser):
     """Set up sys.argv parsers.
 

ecl2df/opmkeywords/WELOPEN

@@ -0,0 +1,37 @@
+{
+  "name": "WELOPEN",
+  "sections": [
+    "SCHEDULE"
+  ],
+  "items": [
+    {
+      "name": "WELL",
+      "value_type": "STRING"
+    },
+    {
+      "name": "STATUS",
+      "value_type": "STRING",
+      "default": "OPEN"
+    },
+    {
+      "name": "I",
+      "value_type": "INT"
+    },
+    {
+      "name": "J",
+      "value_type": "INT"
+    },
+    {
+      "name": "K",
+      "value_type": "INT"
+    },
+    {
+      "name": "C1",
+      "value_type": "INT"
+    },
+    {
+      "name": "C2",
+      "value_type": "INT"
+    }
+  ]
+}

ecl2df/opmkeywords/runmetoupdate.sh

@@ -35,6 +35,7 @@ WCONHIST
 WCONINJE
 WCONINJH
 WCONPROD
+WELOPEN
 WELSEGS
 WELSPECS
 "

tests/test_compdat.py

@@ -181,6 +181,58 @@ def test_tstep():
     assert "2001-05-07" in dates
 
 
+def test_applywelopen():
+    schstr = """
+DATES
+   1 MAY 2001 /
+/
+
+COMPDAT
+ 'OP1' 33 110 31 31 'OPEN'  /
+/
+WELOPEN
+ 'OP1' 'SHUT' /
+/
+
+TSTEP
+  1 /
+
+COMPDAT
+ 'OP2' 66 110 31 31 'OPEN'  /
+/
+
+WELOPEN
+ 'OP1' 'OPEN' /
+/
+
+TSTEP
+  2 3 /
+
+WELOPEN
+ 'OP1' 'POPN' /
+ 'OP2' 'SHUT' /
+/
+"""
+    df = compdat.deck2dfs(EclFiles.str2deck(schstr))["COMPDAT"]
+    assert df.shape[0] == 5
+    assert df["OP/SH"].nunique() == 2
+    assert df["DATE"].nunique() == 3
+
+    schstr = """
+DATES
+   1 MAY 2001 /
+/
+
+COMPDAT
+ 'OP1' 33 110 31 31 'OPEN'  /
+/
+WELOPEN
+ 'OP2' 'SHUT' /
+/"""
+    with pytest.raises(ValueError):
+        compdat.deck2dfs(EclFiles.str2deck(schstr))["COMPDAT"]
+
+
 def test_unrollcompdatk1k2():
     """Test unrolling of k1-k2 ranges in COMPDAT"""
     schstr = """