A Coordinates Transformation Method for Standard Blocks

Posted on Edited on In

1. Introduction

In offshore petroleum exploration practice, the coordinates of blocks designed by geologists for a particular purpose must be provided in Degree-Minute-Second format: the degree uses 2 or 3 digits, the minute uses 2 digits, and the second uses 2 digits, in one of four values — 00, 15, 30 or 45 (in increments of a quarter of a minute). All coordinates must be listed in clockwise order, starting from the name vertex (the most northeast vertex). Such a block is called a “standard block”.

Affleck Oil Field, United Kingdom is an example of such a block.

2. The Needs of Transformation

Designing blocks is daily work in the resource exploration industry. Blocks serve many purposes, including but not limited to administration (relinquishing part of a block, dissolving several blocks into one), block bid campaigns, and so on.

Geologists usually draft blocks on a projected geological setting basemap. Unfortunately, the mapping software in use has no snap functionality. It can only export polygons (the drafted blocks) in projected XY format. It can, however, display a geodetic reference grid at 15-second intervals.

A new workflow was designed to handle the blocks:

  1. Geologists design blocks whose vertices are visually, but not precisely, on the nodes of a 15-second geodetic reference grid.
  2. All blocks reside in a layer, with valid identifiers.
  3. Export the design layer of blocks in projected format as a text file.
  4. Write a program to read the file and generate standard blocks.

3. Implementation

The implementation is a Python package called “transco” (TRANSformation of COordinates), written in 2022 and in use since then.

transco uses pyproj to perform the transformation. pyproj is a Python interface to PROJ. Shapely is also used. Shapely is a Python wrapper for GEOS.

3.1 class

transco uses three classes to represent the data: Layer, Block and Point. Points are properties of a Block instance, and blocks are properties of a Layer instance.

The Layer class is defined as follows:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
class Layer:
    """
    represents a DoubleFox layer contains designed blocks.
    """

    def __init__(
        self,
        layer_file,
        coor_type="proj",
        crs="wgs84utm",
        location="location",
        compute_name=False,
    ):
        self.raw_lines = []
        self.bad_lines = []
        self._blocks = []
        self.cblocks = []
        self.coor_type = coor_type
        self.compute_name = compute_name
        self.name = layer_file.stem.replace("/", "_").replace("\\", "_")
        self.crs = get_crs(coor_type, crs, location)
        self.transformer = None
        self.transform_direction = "proj2geo"
        self.location = location
        self.transformer = get_transformer(self.crs, direction="proj2geo")
        with open(layer_file) as f:
            self.raw_lines = f.readlines()
        lines = copy.deepcopy(self.raw_lines)
        index_for_del = []
        for index, l in enumerate(lines):
            try:
                n, x, y = l.split(",")[:3]
                float(x)
                float(y)
            except:
                index_for_del.append(index)
                self.bad_lines.append(str(index + 1) + "," + l)
                continue
            finally:
                if not n:
                    raise Exception("Found point has no name")
        for i in sorted(index_for_del, reverse=True):
            del lines[i]
        groups = groupby(lines, key=lambda x: x.split(",")[0])
        self.raw_blocks = [
            [p.strip().split(",")[:3] for p in lines] for name, lines in groups
        ]
        for b in self.raw_blocks:
            block = Block(self, b)
            self._blocks.append(block)
        sorted_blocks = sorted(self.blocks, key=lambda x: x.name)
        block_groups = groupby(sorted_blocks, key=lambda x: x.name)
        raw_cblocks = [[b for b in cblock] for name, cblock in block_groups]
        for cb in raw_cblocks:
            _cb = CBlock(self, cb)
            self.cblocks.append(_cb)

    @property
    def blocks(self):
        if len(self._blocks) == len(self.raw_blocks):
            return self._blocks
        raise

    def __str__(self):
        return self.name + ":{} blocks".format(len(self.blocks))

    def __repr__(self):
        return self.__str__()

The coor_type, crs and location properties are used to obtain the corresponding CRS object from pyproj.

The Block class:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
class Block:
    """A raw block resides in a Layer object."""

    def __init__(self, layer, block):
        self.layer = layer
        self.location = self.layer.location
        self.order = 0
        self.crs = self.layer.crs
        self.transformer = self.layer.transformer
        name_str = block[0][0]
        try:
            name, order = name_str.split("--")
            order = int(order)
            if name.strip():
                self.name = name.strip().replace("/", "").replace("\\", "")
            if order >= 1:
                self.order = order
        except:
            self.name = name_str.rstrip("--").strip().replace("/", "").replace("\\", "")
        if not self.name:
            raise Exception("Failed to create a name for this Block")
        self._points = []
        for p in block:
            point = Point(self, p)
            self._points.append(point)
        self.__regular_points()
        self.__unparallel_points()
        self.get_code_name()

    @property
    def points(self):
        return self._points

The block object gets CRS information from its Layer object.

And the Point class looks like:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
class Point:
    """represents a point resides in a Block object."""

    def __init__(self, block, point):
        self.block = block
        self.name = point[0].replace("/", "").replace("\\", "")
        self.x = float(point[1])
        self.y = float(point[2])
        self.lo_deg = None
        self.la_deg = None
        self.lo_dms = None
        self.la_dms = None
        self.offset_lo_dms = None
        self.offset_la_dms = None
        self.crs = self.block.crs
        self.transformer = self.block.transformer
        self._deg()
        self._dms()
        self._offset()
        self._offset_2_dms()

As before, a Point instance gets its CRS from its Block object.

3.2 Function

The basic idea is that the original XY coordinates are transformed into degrees and then snapped to the nearest geodetic reference grid node by the _offset method.

Several methods modify the blocks: rotating points if they are not clockwise, removing redundant points, checking for errors, computing block names by spatial relationship, and so on.

Here is the function for getting CRS.

def get_crs(coor_type, crs, location):
    try:
        epsg_code = defined_crs[coor_type][crs][location]
        return CRS.from_epsg(epsg_code)
    except:
        return None

Transform Degree to DMS:

1
2
3
4
5
6
7
8
9
10
11
12
13
def d2dms(x):
    """
    Transform a single float degree value to DMS format.
    return a tuple of degree, minute, second in float.
    >>> d2dms(113.5)
    (113.0, 30.0, 0.0)
    >>> d2dms(20.254166667)
    (20.0, 15.0, 15.000001199999247)
    """
    d_decimal, d = math.modf(x)
    m_decimal, m = math.modf(d_decimal * 60)
    s = m_decimal * 60
    return d, m, s

The last step is defining functions to call methods on the objects.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
def load_file(
    file="df_design.txt", coor_type="proj", crs="wgs84utm", location="location"
):
    f = Path(file)
    return Layer(f, coor_type, crs, location)

def save_default(layer):
    layer.save_standardized_blocks()
    layer.save_blocks_area_info_wgs84utm()

def available_crs(defined_crs):
    all_crs = {}
    for key, value in defined_crs.items():
        if value:
            all_crs[key] = value
    return all_crs

def show_help(**kw):
    print(__desc__)
    print("Version {}".format(__version__))
    print("Author: ", __author__)
    print(__transco_help__)
    location = kw.get("location", None)
    if location:
        try:
            crs = get_crs("proj", "wgs84utm", location)
        except:
            print("Failed to get CRS.")
            raise
        if crs:
            print("Current CRS:")
            print(crs.name)
            print("Area of Use:")
            print(crs.area_of_use)

def main():
    parser = argparse.ArgumentParser(description=__desc__)
    parser.add_argument(
        "file", help="the input file, default is 'df_design.txt'.", nargs="?"
    )
    parser.add_argument(
        "location",
        help="the location for projection selection, default is 'location'.",
        nargs="?",
    )
    parser.add_argument(
        "-a", "--all", action="store_true", help="save all output files"
    )
    parser.add_argument("--version", action="store_true", help="show version")
    parser.add_argument("--allcrs", action="store_true", help="show all available CRS")
    args = parser.parse_args()
    try:
...

After that, I can call the main function from the CLI, call the functions and classes directly in an interactive Python environment, or use it in a Jupyter notebook.

4. Results

transco has been in use since 2022. A 50-block batch that previously took roughly an hour per block to prepare by hand in Excel now runs in about 10 minutes.

The larger gain is in the revision loop. Geologists revise their block designs repeatedly before a final version is agreed, and the manual cost was paid again on every revision. Re-running transco costs nothing, so a design can be iterated as many times as the work requires.