more consistent error handling

errs.go

@@ -0,0 +1,44 @@
+package grate
+
+import "errors"
+
+var (
+	// configure at build time by adding go build arguments:
+	//   -ldflags="-X github.com/pbnjay/grate.loglevel=debug"
+	loglevel string = "warn"
+
+	// Debug should be set to true to expose detailed logging.
+	Debug bool = (loglevel == "debug")
+)
+
+// ErrInvalidScanType is returned by Scan for invalid arguments.
+var ErrInvalidScanType = errors.New("grate: Scan only supports *bool, *int, *float64, *string, *time.Time arguments")
+
+// ErrNotInFormat is used to auto-detect file types using the defined OpenFunc
+// It is returned by OpenFunc when the code does not detect correct file formats.
+var ErrNotInFormat = errors.New("grate: file is not in this format")
+
+// ErrUnknownFormat is used when grate does not know how to open a file format.
+var ErrUnknownFormat = errors.New("grate: file format is not known/supported")
+
+type errx struct {
+	errs []error
+}
+
+func (e errx) Error() string {
+	return e.errs[0].Error()
+}
+func (e errx) Unwrap() error {
+	if len(e.errs) > 1 {
+		return e.errs[1]
+	}
+	return nil
+}
+
+// WrapErr wraps a set of errors.
+func WrapErr(e ...error) error {
+	if len(e) == 1 {
+		return e[0]
+	}
+	return errx{errs: e}
+}

go.mod

@@ -0,0 +1,3 @@
+module github.com/pbnjay/grate
+
+go 1.16

grate.go

@@ -4,6 +4,7 @@ package grate
 
 import (
 	"errors"
+	"log"
 	"sort"
 )
 
@@ -16,14 +17,32 @@ type Source interface {
 	Get(name string) (Collection, error)
 }
 
+// Collection represents an iterable collection of records.
+type Collection interface {
+	// Next advances to the next record of content.
+	// It MUST be called prior to any Scan().
+	Next() bool
+
+	// Strings extracts values from the current record into a list of strings.
+	Strings() []string
+
+	// Scan extracts values from the current record into the provided arguments
+	// Arguments must be pointers to one of 5 supported types:
+	//     bool, int, float64, string, or time.Time
+	// If invalid, returns ErrInvalidScanType
+	Scan(args ...interface{}) error
+
+	// IsEmpty returns true if there are no data values.
+	IsEmpty() bool
+
+	// Err returns the last error that occured.
+	Err() error
+}
+
 // OpenFunc defines a Source's instantiation function.
 // It should return ErrNotInFormat immediately if filename is not of the correct file type.
 type OpenFunc func(filename string) (Source, error)
 
-// ErrNotInFormat is used to auto-detect file types using the defined OpenFunc
-// It is returned by OpenFunc when the code does not detect correct file formats.
-var ErrNotInFormat = errors.New("grate: file is not in this format")
-
 // Open a tabular data file and return a Source for accessing it's contents.
 func Open(filename string) (Source, error) {
 	for _, o := range srcTable {
@@ -31,11 +50,14 @@ func Open(filename string) (Source, error) {
 		if err == nil {
 			return src, nil
 		}
-		if err != ErrNotInFormat {
+		if !errors.Is(err, ErrNotInFormat) {
 			return nil, err
 		}
+		if Debug {
+			log.Println(" ", filename, "is not in", o.name, "format")
+		}
 	}
-	return nil, errors.New("grate: file format is not known/supported")
+	return nil, ErrUnknownFormat
 }
 
 type srcOpenTab struct {
@@ -54,24 +76,3 @@ func Register(name string, priority int, opener OpenFunc) error {
 	})
 	return nil
 }
-
-// Collection represents an iterable collection of records.
-type Collection interface {
-	// Next advances to the next record of content.
-	// It MUST be called prior to any Scan().
-	Next() bool
-
-	// Strings extracts values from the current record into a list of strings.
-	Strings() []string
-
-	// Scan extracts values from the current record into the provided arguments
-	// Arguments must be pointers to one of 5 supported types:
-	//     bool, int, float64, string, or time.Time
-	Scan(args ...interface{}) error
-
-	// IsEmpty returns true if there are no data values.
-	IsEmpty() bool
-
-	// Err returns the last error that occured.
-	Err() error
-}

simple/csv.go

@@ -33,6 +33,17 @@ func OpenCSV(filename string) (grate.Source, error) {
 		total++
 		t.rows = append(t.rows, rec)
 	}
+	if err != nil {
+		switch perr := err.(type) {
+		case *csv.ParseError:
+			return nil, grate.WrapErr(perr, grate.ErrNotInFormat)
+		}
+		if total < 10 {
+			// probably? not in this format
+			return nil, grate.WrapErr(err, grate.ErrNotInFormat)
+		}
+		return nil, err
+	}
 
 	// kinda arbitrary metrics for detecting CSV
 	looksGood := 0

simple/simple.go

@@ -67,9 +67,9 @@ func (t *simpleFile) Scan(args ...interface{}) error {
 		case *string:
 			*v = row[i]
 		case *time.Time:
-			return errors.New("grate/simple: time.Time not supported, you must parse string manually")
+			return errors.New("grate/simple: time.Time not supported, you must parse date strings manually")
 		default:
-			return errors.New("grate/simple: scan destination must be one of: *bool, *int, *float64, *string, or *time.Time")
+			return grate.ErrInvalidScanType
 		}
 		if err != nil {
 			return err

simple/tsv.go

@@ -32,6 +32,10 @@ func OpenTSV(filename string) (grate.Source, error) {
 		total++
 		t.rows = append(t.rows, r)
 	}
+	if s.Err() != nil {
+		// this can only be read errors, not format
+		return nil, s.Err()
+	}
 
 	// kinda arbitrary metrics for detecting TSV
 	looksGood := 0

xls/cfb/interface.go

@@ -4,6 +4,8 @@ import (
 	"fmt"
 	"io"
 	"os"
+
+	"github.com/pbnjay/grate"
 )
 
 // Document represents a Compound File Binary Format document.
@@ -24,7 +26,7 @@ func Open(filename string) (Document, error) {
 	}
 	err = d.load(f)
 	if err != nil {
-		return nil, err
+		return nil, grate.WrapErr(err, grate.ErrNotInFormat)
 	}
 	return d, nil
 }

xls/sheets.go

@@ -503,15 +503,12 @@ func (s *WorkSheet) Scan(args ...interface{}) error {
 		case *time.Time:
 			*v = currow.cols[i].(time.Time)
 		default:
-			return ErrInvalidType
+			return grate.ErrInvalidScanType
 		}
 	}
 	return nil
 }
 
-// ErrInvalidType is returned by Scan for invalid arguments.
-var ErrInvalidType = errors.New("xls: Scan only supports *bool, *int, *float64, *string, *time.Time arguments")
-
 var berrLookup = map[byte]string{
 	0x00: "#NULL!",
 	0x07: "#DIV/0!",

xls/xls.go

@@ -51,7 +51,7 @@ func (b *WorkBook) IsProtected() bool {
 func Open(filename string) (grate.Source, error) {
 	doc, err := cfb.Open(filename)
 	if err != nil {
-		return nil, grate.ErrNotInFormat //err
+		return nil, err
 	}
 
 	b := &WorkBook{
@@ -64,7 +64,7 @@ func Open(filename string) (grate.Source, error) {
 
 	rdr, err := doc.Open("Workbook")
 	if err != nil {
-		return nil, err
+		return nil, grate.WrapErr(err, grate.ErrNotInFormat)
 	}
 	err = b.loadFromStream(rdr)
 	return b, err

xlsx/sheets.go

@@ -252,7 +252,7 @@ func (s *Sheet) Scan(args ...interface{}) error {
 		case *time.Time:
 			*v = currow.cols[i].(time.Time)
 		default:
-			return ErrInvalidType
+			return grate.ErrInvalidScanType
 		}
 	}
 	return nil
@@ -266,6 +266,3 @@ func (s *Sheet) IsEmpty() bool {
 func (s *Sheet) Err() error {
 	return s.err
 }
-
-// ErrInvalidType is returned by Scan for invalid arguments.
-var ErrInvalidType = errors.New("xlsx: Scan only supports *bool, *int, *float64, *string, *time.Time arguments")

xlsx/xlsx.go

@@ -40,7 +40,7 @@ func Open(filename string) (grate.Source, error) {
 	}
 	z, err := zip.NewReader(f, info.Size())
 	if err != nil {
-		return nil, grate.ErrNotInFormat //err
+		return nil, grate.WrapErr(err, grate.ErrNotInFormat)
 	}
 	d := &Document{
 		filename: filename,
@@ -52,12 +52,12 @@ func Open(filename string) (grate.Source, error) {
 	// parse the primary relationships
 	dec, c, err := d.openXML("_rels/.rels")
 	if err != nil {
-		return nil, err
+		return nil, grate.WrapErr(err, grate.ErrNotInFormat)
 	}
 	err = d.parseRels(dec, "")
 	c.Close()
 	if err != nil {
-		return nil, err
+		return nil, grate.WrapErr(err, grate.ErrNotInFormat)
 	}
 	if d.primaryDoc == "" {
 		return nil, errors.New("xlsx: invalid document")